Validation of Thermal Destruction of Pathogenic Bacteria ...

Clemson UniversityTigerPrints

All Dissertations Dissertations

12-2013

Validation of Thermal Destruction of PathogenicBacteria in Rendered Animal ProductsMary HayesClemson University, [email protected]

Follow this and additional works at: https://tigerprints.clemson.edu/all_dissertations

Part of the Microbiology Commons

This Dissertation is brought to you for free and open access by the Dissertations at TigerPrints. It has been accepted for inclusion in All Dissertations byan authorized administrator of TigerPrints. For more information, please contact [email protected].

Recommended CitationHayes, Mary, "Validation of Thermal Destruction of Pathogenic Bacteria in Rendered Animal Products" (2013). All Dissertations. 1255.https://tigerprints.clemson.edu/all_dissertations/1255

https://tigerprints.clemson.edu?utm_source=tigerprints.clemson.edu%2Fall_dissertations%2F1255&utm_medium=PDF&utm_campaign=PDFCoverPages

https://tigerprints.clemson.edu/all_dissertations?utm_source=tigerprints.clemson.edu%2Fall_dissertations%2F1255&utm_medium=PDF&utm_campaign=PDFCoverPages

https://tigerprints.clemson.edu/dissertations?utm_source=tigerprints.clemson.edu%2Fall_dissertations%2F1255&utm_medium=PDF&utm_campaign=PDFCoverPages

https://tigerprints.clemson.edu/all_dissertations?utm_source=tigerprints.clemson.edu%2Fall_dissertations%2F1255&utm_medium=PDF&utm_campaign=PDFCoverPages

http://network.bepress.com/hgg/discipline/48?utm_source=tigerprints.clemson.edu%2Fall_dissertations%2F1255&utm_medium=PDF&utm_campaign=PDFCoverPages

https://tigerprints.clemson.edu/all_dissertations/1255?utm_source=tigerprints.clemson.edu%2Fall_dissertations%2F1255&utm_medium=PDF&utm_campaign=PDFCoverPages

mailto:[email protected]

VALIDATION OF THERMAL DESTRUCTION OF PATHOGENIC BACTERIA IN

RENDERED ANIMAL PRODUCTS

A Dissertation

Presented to

the Graduate School of

Clemson University

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

Microbiology

by

Mary Melissa Hayes

December 2013

Accepted by:

Annel K. Greene, Committee Chair

Thomas A. Hughes

Xuiping Jiang

Paul L. Dawson

ii

ABSTRACT

This study tested enumeration techniques for high fat food matrices and

determined thermal death times in rendering animal products. Standard Class O

phosphate/magnesium chloride dilution buffer series (Dilution Series A) and a modified

(pre-warmed to 32ºC) lecithin phosphate dilution buffer series (Dilution Series B) were

used to enumerate a Salmonella cocktail from both poultry and beef rendering materials.

Results of this study indicate use of a modified lecithin buffer did not improve

Salmonella enumeration accuracy from rendering materials. Instead, the results suggested

use of xylose lysine deoxycholate agar (XLD) with either buffer system accurately

enumerated Salmonella from rendering materials.

The thermal death of four pathogenic strains of Salmonella recognized by the

FDA as hazardous in animal feeds (Salmonella Choleraesuis (SC), Salmonella Enteriditis

(SE), Salmonella Newport (SN), and Salmonella Dublin (SD)) was not a straight line

decrease. After periods of appearing to be destroyed, some cultures reappeared at later

treatment times. In thermal treatments up to 420 s at 240ºF (115.6ºC), SC was last

detected at 120 s, SE at 120 s, SN at 300 s and SD at 360 s in inoculated beef rendering

materials. In thermal treatments up to 420 s at 240ºF (115.6ºC), SC, SE, SN, and SD were

last detected at 360 s, respectively, in inoculated poultry rendering materials. Controls

indicated thermally resistant strains in the background of both beef and poultry rendering

materials which when tested using standard FDA Bacteriological Analytical Manual

(BAM) techniques indicated Salmonella. Hypotheses to explain the results of this study

include: 1) thermally resistant sub-particles such as bone or tissue protected bacteria from

iii

thermal treatment; 2) presence of thermally resistant species in the background of

rendering samples caused false positive results on BAM procedures; or 3) presence of

thermally resistant Salmonella. Further research will need to be conducted at 240ºF

(115.6ºC) for longer time intervals to ensure that SC, SE, SN and SD are destroyed and to

identify the impact of particles on thermal conductivity through the rendering matrices.

Additionally, future experimentation will be needed to verify that the microorganisms

identified are indeed Salmonella or other another microorganism(s) cross-reacting as

Salmonella.

iv

DEDICATION

This dissertation is dedicated to my family and many friends. A special dedication

in gratitude goes to my grandmother, Ellen Davis, whose words of encouragement helped

me to achieve my goals. I dedicate this work to my parents, Ann and Michael Hayes, and

my brother, Michael Hayes. Without their love and support, this could not have been

possible. I also dedicate this work and give special thanks to my fiancé, Brian Gaines, for

being there for me on the tough days. Finally, I would like to dedicate my dissertation to

Neko, Bella, Vixen, Hitch, Nokie, Kitty, and Willow, for their hugs and unconditional

love.

v

ACKNOWLEDGMENTS

I acknowledge, with much gratitude, my debt of thanks to Dr. Annel K. Greene

for her advice, support, and foresight. I also would like to thank my committee members,

Dr. Thomas A. Hughes, Dr. Xiuping Jiang, and Dr. Paul L. Dawson for their constant

guidance during my research project. I am eternally grateful to all the undergraduate

assistants for helping with all the tedious grunt work in the laboratory. Finally, I would

like to thank Steven D. Chambers and Yubo Zhang for being a surrogate family offering

continued moral support in and outside of the laboratory.

vi

TABLE OF CONTENTS

Page

TITLE PAGE .................................................................................................................... i

ABSTRACT ..................................................................................................................... ii

DEDICATION ................................................................................................................ iv

ACKNOWLEDGMENTS ............................................................................................... v

LIST OF TABLES ........................................................................................................ viii

LIST OF FIGURES ......................................................................................................... x

CHAPTER

1. LITERATURE REVIEW .............................................................................. 1

Introduction .............................................................................................. 1

The Rendering Process ............................................................................ 2

Thermal Processing Principles ................................................................. 6

Salmonella................................................................................................ 8

Salmonella and Rendering Co-Products ................................................ 10

Emulsifiers ............................................................................................. 13

Conclusion ............................................................................................. 14

References .............................................................................................. 15

2. METHOD VALIDATION FOR ENUMERATING BACTERIA IN HIGH

FAT MATRICES ................................................................................... 21

Abstract .................................................................................................. 21

Introduction ............................................................................................ 22

Materials and Methods ........................................................................... 23

Rendering Sample Preparation .................................................. 23

Salmonella Preparation .............................................................. 24

Salmonella Enumeration in Rendering Materials ...................... 25

Statistical Analysis ..................................................................... 27

Results .................................................................................................... 27

Discussion .............................................................................................. 29

vii

Table of Contents (Continued)

Page

Figures and Tables ................................................................................. 32

References .............................................................................................. 34

3. VALIDATION OF THERMAL DESTRUCTION OF SALMONELLA IN

RENDERED BEEF PRODUCTS.......................................................... 35

Abstract .................................................................................................. 35

Introduction ............................................................................................ 37




Thermal Death Time Trials ........................................................ 41

Bone Particle Size Determination .............................................. 44

Determination of Estimated D Values ....................................... 44

Results .................................................................................................... 45

Discussion .............................................................................................. 49


References .............................................................................................. 71

4. VALIDATION OF THERMAL DESTRUCTION OF SALMONELLA IN

RENDERED POULTRY PRODUCTS ................................................. 73

Abstract .................................................................................................. 73

Introduction ............................................................................................ 74




Thermal Death Time Trials ........................................................ 79

Determination of Estimated D Values ....................................... 81

Results .................................................................................................... 82

Discussion .............................................................................................. 85


References ............................................................................................ 105

viii

LIST OF TABLES

Table Page

2.1 Comparison of mean bacterial counts from each poultry

and beef rendering material using each dilution

series to a standard culture control......................................................... 33

3.1 Salmonella Choleraesuis (SC), Salmonella Enteriditis

(SE), Salmonella Newport (SN) and Salmonella

Dublin (SD) after inoculation into beef rendering

materials and plated onto XLD .............................................................. 64

3.2 Estimated D values for Salmonella Choleraesuis (SC),

Salmonella Enteriditis (SE), Salmonella Newport

(SN), and Salmonella Dublin (SD) in beef rendering

samples (50% fat) at 115.6ºC pre-enriched in RV

and validated by two confirmation tests ................................................ 65



(SN), and Salmonella Dublin (SD) in beef rendering

samples (50% fat) at 115.6ºC pre-enriched in TT

and validated by two confirmation tests ................................................ 66

3.4 Number of samples positive for Salmonella in

Salmonella Choleraesuis (SC), Salmonella

Enteriditis (SE), Salmonella Newport (SN)

and Salmonella Dublin (SD) inoculated beef

rendering samples (50% fat) after pre-enrichment

in RV or TT broth .................................................................................. 67


uninoculated beef rendering samples

(50% fat) for Salmonella Choleraesuis (SC),


(SN) and Salmonella Dublin (SD) after pre-

enrichment in RV or TT broth ............................................................... 69

4.1 Salmonella Choleraesuis (SC), Salmonella

Enteriditis (SE), Salmonella Newport (SN) and

Salmonella Dublin (SD) after inoculation into

poultry rendering materials and plated onto XLD ................................. 98

ix

List of Tables (Continued)

Table Page



(SN), and Salmonella Dublin (SD) in poultry

rendering samples (50% fat) at 115.6ºC pre-enriched

in RV and validated by two confirmation tests ...................................... 99



(SN), and Salmonella Dublin (SD) in poultry

rendering samples (50% fat) at 115.6ºC pre-enriched

in TT and validated by two confirmation tests .................................... 100

4.4 Number of samples positive for Salmonella in Salmonella

Choleraesuis (SC), Salmonella Enteriditis (SE),

Salmonella Newport (SN) and Salmonella Dublin

(SD) inoculated poultry rendering samples (50% fat)

after pre-enrichment in RV or TT broth .............................................. 101


uninoculated poultry rendering samples (50% fat)

for Salmonella Choleraesuis (SC), Salmonella

Enteriditis (SE), Salmonella Newport (SN) and

Salmonella Dublin (SD) after pre-enrichment in RV

or TT broth ........................................................................................... 103

x

LIST OF FIGURES

Figure Page

2.1 Diagram of two buffer systems used for serially diluting

rendering materials................................................................................. 32

3.1 Method utilized to report RV and TT pre-enrichments

results on XLD validated by two confirmation

tests at each thermal treatment ............................................................... 54

3.2 Enumeration of Salmonella on XLD from beef rendering

samples (50% fat) inoculated with Salmonella


Salmonella Newport (SN), and Salmonella Dublin

(SD) ........................................................................................................ 55

3.3 Enumeration of Salmonella on XLD from uninoculated

beef rendering samples (50% fat) for Salmonella



(SD) ........................................................................................................ 56

3.4 Enumeration of total bacteria on TSA from beef rendering




(SD) ........................................................................................................ 57

3.5 Enumeration of total bacteria on TSA from uninoculated

beef rendering samples (50% fat) for Salmonella



(SD) ........................................................................................................ 58

3.6 Presence or absence ± standard deviation of Salmonella

for each Salmonella Choleraesuis (SC), Salmonella

Enteriditis (SE), Salmonella Newport (SN), and

Salmonella Dublin (SD) inoculated, RV pre-enriched

beef rendering samples (50% fat) .......................................................... 59

xi

List of Figures (Continued)

Figure Page


for each RV pre-enriched, uninoculated beef rendering

samples (50% fat) for Salmonella Choleraesuis (SC),

Salmonella Enteriditis (SE), Salmonella Newport (SN),

and Salmonella Dublin (SD) .................................................................. 60




Salmonella Dublin (SD) inoculated, TT pre-enriched

beef rendering samples (50% fat) .......................................................... 61


for each TT pre-enriched, uninoculated beef rendering




3.10 Mean percent particle size distribution ± standard error

of beef rendering samples collected from a rendering

plant on three different days .................................................................. 63

4.1 Method utilized to report RV and TT pre-enrichments

results on XLD confirmed by two confirmation tests

at each thermal treatment ....................................................................... 89

4.2 Enumeration of Salmonella on XLD from poultry rendering




(SD) ........................................................................................................ 90

4.3 Enumeration of Salmonella on XLD from uninoculated

poultry rendering samples (50% fat) for Salmonella



(SD) ........................................................................................................ 91

xii

List of Figures (Continued)

Figure Page

4.4 Enumeration of total bacteria on TSA from poultry rendering




(SD) ........................................................................................................ 92

4.5 Enumeration of total bacteria on TSA from uninoculated

poultry rendering samples (50% fat) for Salmonella



(SD) ........................................................................................................ 93




Salmonella Dublin (SD) inoculated, RV pre-enriched

poultry rendering samples (50% fat) ...................................................... 94


for each RV pre-enriched, uninoculated poultry rendering



and Salmonella Dublin (SD). ................................................................. 95




Salmonella Dublin (SD) inoculated, TT pre-enriched

poultry rendering samples (50% fat) ...................................................... 96


for each TT pre-enriched, uninoculated poultry rendering




1

CHAPTER 1

LITERATURE REVIEW

Introduction

Rendering is the recycling of residual animal tissue from food animals into stable,

value-added materials for animal feeds, chemical feedstocks, and fertilizers.

Approximately 50% of a food animal is considered edible; the remainder of the animal

tissue is rendered into animal co-products (Meeker and Hamilton 2006). Non-utilization

of animal co-products would create major aesthetic and potentially disastrous public

health problems since these organic materials are highly perishable and laden with

microorganisms, many of which can cause disease in both humans and animals (Meeker

and Hamilton 2006).

Approximately 8 billion chickens, 1.6 billion turkeys, 100 million hogs, and 35

million cattle are slaughtered and processed each year in the United States (Meeker and

Hamilton 2006; Richardson 2006). However, on average, only 51% of the live weight of

cattle, 56% of the live weight of hogs, 63% of the live weight of broilers, and 43% of the

live weight of most fish species can be considered edible by Americans and Canadians

(Meeker and Hamilton 2006). Due to various dietary practices and taste preferences

around the world, the term “edible” may be construed in different ways depending on the

region or country. However, non-carcass materials such as liver, tongue, heart, kidney,

thymus, stomach, cheeks, head trimmings, blood, lungs, fat, and bones are a source of

nutrients and can be consumed by animals if properly processed by the rendering industry

(Ockerman and Hansen 2000; Dos Santos 2013). In addition to non-carcass material,

2

increasing amounts of waste ready-to-eat (RTE) and/or heat-and-eat packaged foods are

processed by the rendering industry (Kinley 2009). The rendering industry also processes

waste cooking fats and oils from restaurants. The United States and Canadian rendering

industry annually recycles over 61 billion pounds of residual products into animal feeds,

fats and proteins to prevent waste of these materials and the overfilling of landfills. The

rendering facilities predominantly produce meat and bone meal, poultry meal, hydrolyzed

feather meal, blood meal, fish meal, and animal fats. In the United States, the rendering

industry annually produces approximately 11.2 billion pounds of protein and 10.9 billion

pounds of fats. Approximately 85% of rendered products are used as animal feed

ingredients for livestock and pets (Meeker and Hamilton 2006). However, the National

Renderers Association has reported over 3000 rendering product industrial applications

identified in many areas including personal care, biofuel, and chemical industries

(Meeker and Hamilton 2006).

Certain provisions are necessary for animal co-products to be effectively used.

These requirements include a sufficient volume of animal co-products in a centralized

location, a method to commercially process animal co-products into marketable goods, an

efficient market to sell products produced from animal co-products, and storage systems

for finished animal co-products. Not meeting these requirements leads to under-

utilization of animal co-products (Ockerman and Hansen 2000; Clemen 1927).

The Rendering Process

Rendering is a process that involves heat and other procedures to separate water,

fat, and protein contained in animal tissues. The temperature and length of time of the

3

cooking process can impact the quality of the finished product. Despite the type of raw

material being processed, the rendering process is comprised of several universal phases

(ICMSF 2000; Kinley 2009). Initially all raw material is transported to an area in the

rendering plant where, if necessary, it is pre-crushed to reduce size prior to being fed onto

a conveyer that transfers it into the cooker. The basic process involves collecting and

sizing raw material as needed, heating to remove the water, removing the fat by draining

and/or pressing, cooling, milling, and storing. Many variations of these operations have

been developed according to the type of raw material, machinery, and the facility

(ICMSF 2000). Raw animal materials vary but these materials typically contain

approximately 60% water, 20% protein and mineral, and 20% fat before the rendering

process (Meeker and Hamilton 2006).

Without barriers and other protections, the aerosols generated during raw material

crushing have the potential to spread contaminating microbes in the rendering facility

including areas where the finished product is handled (ICMSF 2000; Swingler 1982). The

rendering cooking process is reported to be 40 to 90 min at 240 to 290ºF (115.6 to

143.3ºC) (Meeker and Hamilton 2006). Process control is performed and monitored via

computers so that time/temperature processes for appropriate moisture loss is achieved.

However, the exact time and temperature relationship for thermal death of specific

microorganisms has not been established in rendering matrices. It has been demonstrated

that the high fat and low water environment of batch dry rendered material will protect

bacterial spores against thermal inactivation (Lowry et al. 1979; ICMSF 2000).

4

Therefore, some marginal processing conditions could result in finished products

containing bacterial spores or other heat-resistant biological materials.

In North America, most of the rendering facilities utilize continuous-flow, dry

rendering units. Only large animals such as cattle and hogs are crushed or chopped prior

to processing, whereas smaller animals such as poultry are not ground prior to cooking.

Once in the continuous cooker, steam is utilized to heat the internal metal components of

the rendering cooker. In this type of dry continuous cooker unit, the steam transfers heat

across metal heating surfaces to the rendering materials. The steam is condensed in a

closed loop system so the water will never come in contact with the rendering materials.

The condensed water is transported out of the cooker back to the steam generator

(Ockerman and Hansen 2000; Kinley 2009).

In order to thermally process in a rendering cooker, raw materials are deposited

into hot rendered fat and during the cooking process, moisture is removed by

evaporation. After the cooking process, the protein/bone material and molten fat are

initially separated by a screen drainer and an auger conveyor that moves the materials to

the screw press. The screw press removes additional fat content from solid material

(Ockerman and Hansen 2000; Anderson 2006; Kinley 2009). The remaining material

known as “cracklings” or “crax” is ground (Ockerman and Hansen 2000). Both the

ground processed protein meal and fat are transferred to a storage facility or transported

to a consumer (Meeker and Hamilton, 2006). The processed protein meal is either stored

in feed bin structures or enclosed silos. Fat is centrifuged to remove residual particulate

and stored in insulated and/or heated silos. The renderers maintain the fat at elevated

5

temperatures to keep it in a liquid form in order to easily pump the fat from the silos.

Despite the elevated temperatures utilized during the rendering process, the finished

products are susceptible to recontamination from raw materials and the rendering facility

environment. In 2000, microbial levels in finished rendered products were reportedly

high (ICMSF 2000) but the rendering industry is continuing to make major improvements

to reduce microbial levels in finished products. Educational programs such as the Code of

Practice Seminar initiated in 2004, use of HAACP, and certification are offered through

the American Protein Producers Industry (APPI) audit program to improve

microbiological quality of rendered products (Meeker and Hamilton, 2006).

Wet and batch rendering units are used in North American rendering plants in

addition to continuous flow dry rendering units (Ockerman and Hansen 2000; Anderson

2006). In a wet rendering facility, steam is injected directly in contact with the product by

vertical digesters through perforated plates, which can produce high quality tallow. These

inefficient systems are labor intensive, require long cooking times, lose large volumes of

meal during processing, and produce high moisture products (Ockerman and Hansen

2000). Batch rendering systems have expensive operation costs and are unable to quickly

process large volumes of materials continuously (Ockerman and Hansen 2000).

Continuous slurry systems such as the Carver-Greenfield system are utilized in some

rendering facilities. These systems produce a more digestible meal and high quality fat.

Continuous slurry systems are energy efficient, however, they process at temperatures

close to 240ºF (115.6ºC) (Meeker and Hamilton, 2006) thereby possibly not effective at

destroying some bacterial species.

6

Thermal Processing Principles

When microbial populations in food or rendering materials are exposed to

elevated temperatures, the microbial population reduction typically occurs in a

logarithmic (log10) manner with increasing time at a given constant elevated temperature.

Several parameters are utilized to quantify the influence of elevated temperatures on

microbial populations. Thermal death time (TDT) or F value is a factor of time,

temperature, material matrix and organism (Heldman and Hartel 1998). TDT is defined

as the time needed to kill or reduce a given number of organisms at a specific temperature

(Jay 2005; Teixeira 2006). TDT can be utilized as a measure of product safety to reduce a

microbial population in a product to decrease spoilage microbes and increase shelf-life.

Decimal reduction time or D value indicates the time required for a one log10 cycle

reduction of a particular organism at a specific temperature. Essentially, a large D value

at a given temperature indicates an increased thermal resistance of a microbial population

in a product (Heldman and Hartel 1998). The 12-D concept is used as a lethality time

required for the canning industry and is defined as the time required for destroying 12

log10 of Clostridium botulinum spores (Jay 2005; Teixeira 2006). The thermal resistance

constant or Z value is the parameter used to indicate the temperature increase needed to

cause a one log10 reduction as shown as the slope on the thermal destruction curve. In

most situations, a large Z value would indicate that a microbial culture contains heat

resistant vegetative cells or microbial spores (Heldman and Hartel 1998).

Numerous research studies have been conducted in the food industry regarding

different factors such as cooking methods, food composition, packaging type and product

7

type and their impact on the thermal lethality of pathogens. Blackburn et al. (1997)

developed and validated thermal inactivation models for Salmonella Enteriditis and

Escherichia coli O157:H7 describing the effect of temperature, pH and sodium chloride

concentration on each microbe in whole egg, egg albumen, egg yolk, beef, poultry, apple

juice and milk. Orta-Ramirez et al. (1997) demonstrated the temperature dependence of

the enzyme triose phosphate isomerase from E. coli O157:H7 and Salmonella senftenberg

indicating this enzyme could potentially be used as a surrogate time-temperature

indicator in ground beef products. Juneja et al. (2000) determined beef samples

containing between 7 and 24% fat content and inoculated with a Salmonella cocktail had

varying D and Z values. Murphy et al. (2002) and Murphy et al. (2004) reported

Salmonella and Listeria innocua had significantly different thermal inactivation D and Z

values among several different commercial products such as chicken breast meat, chicken

patties, chicken tenders, franks, beef patties, blended beef and turkey patties.

Although human food products and their processing systems provide the nearest

similarities for studying microbiological population dynamics inherent in rendered

animal products, there is not an ideal model found in the food industry to duplicate

rendered materials. Procedures used for food microbiological testing are unproven in

rendered animal product testing. For instance, after attempting to quantify microbial

loads in raw poultry rendering materials, Glenn (2006) discovered difficulties in

enumerating bacteria by traditional aqueous buffer dilution methods due to the high fat

content of the rendering material. Rendered animal co-products are a combination of

various offal tissues, bones and fat (Meeker and Hamilton, 2006), and these materials

8

have unique compositions not found in any known food product. Therefore, the high fat,

bone and protein content of rendering materials leaves the industry with no comparable

thermal death time values from the human food industry or any other industry. Since

thermal death time is a factor of matrix, temperature and organism, it will be necessary to

conduct validation in the actual rendering material matrices. The high fat content of

rendered products also complicates traditional bacterial enumeration methodology. It is

imperative that accurate test methods are developed to detect these pathogens in high fat

rendered materials to prevent false positive and false negative results.

Salmonella

Salmonella is a genus of Gram-negative, facultative anaerobic, motile, non-spore-

forming bacilli which are classified as members of the Enterobacteriaceae family (Wray

and Wray 2000). Certain species of Salmonella are associated with foodborne disease

(Jay 2005). Typically, Salmonella are characterized by their ability to ferment glucose

into gas and acid on triple sugar iron (TSI) media and will not utilize sucrose or lactose in

differential media (Andrews et al. 2011; D’Aoust et al. 1998). However, in some cases,

Salmonella have demonstrated ability to ferment sucrose and lactose through the use of

plasmids (Le Minor et al. 1973; Le Minor et al. 1974).

The optimal growth temperature for Salmonella is 37ºC and growth is faster in

moist conditions (Franco 1997). These organisms are able to multiply over a wide variety

of conditions including extreme temperatures (high and low) and low water activity

levels (Franco 1997). Some strains of Salmonella have been able to grow in environments

as high as 54ºC and some as low as 2ºC (D’Aoust et al. 1975). Salmonella can develop

9

heat resistance by exposure of the cells to temperatures greater than 50ºC between 15 and

30 min resulting in production of heat shock proteins (Humphrey et al. 1993; Mackey and

Derrick 1986; Mackey and Derrick 1990).

Rasmussen et al. (1964) reported wet Salmonella cells added to dry meat and

bone meal were reduced by 8 log10 after the meal was heated for 15 min at 68ºC;

however, heating for 1 h at 82ºC was required to kill Salmonella cells in naturally

contaminated meal. The water activity was not reported in this study. Mossel et al. (1965)

observed a rapid 5 log10 reduction immediately after inoculation of a viable Salmonella

culture containing 10 to 12 log10 concentration of cells. The broth culture was pre-chilled

at 4ºC and mixed into dry meat and bone meal which also was pre-chilled at 4ºC. The

water activity level of the meat and bone meal was reported as 0.46. Mossel et al. (1965)

theorized that the initial rapid decline of the Salmonella concentration was due to osmotic

shock. After additional storage for 5 days under refrigeration temperatures, a 1 log10

reduction of the Salmonella culture occurred. Mossel et al. (1965) noted that once

bacterial cells are within protein protected by lipids, increased resistance seemed to

occur. Reinman (1968) indicated a drastic reduction in viable Salmonella after meat and

bone meal (water activity of 0.9) was heated to 90ºC for a relatively short time.

Genetic mutations in strains of Salmonella also can increase heat resistance.

Droffner and Yamamoto (1992) determined Salmonella Typhimurium was capable of

surviving prolonged exposure at 54ºC. The results of this study indicated genetic

mutations occurred in the ttl gene or the mth gene which gives increased heat resistance

at temperatures as high as 48ºC and 54ºC, respectively. In addition to these genes, other

10

environmental factors play a role in the level of heat resistance including the nutrients

available in the growth environment, the growth phase of the cells, and the moisture level

of the environment from which it was isolated (Goepfert et al. 1970; Kirby and Davies

1990; Ng et al. 1969).

Salmonella and Rendering Co-Products

In the United States, approximately 2 to 4 million cases of human salmonellosis

occur annually (FDA 2012). Often these Salmonella outbreaks are associated with

consumption of animal products (Shacher and Yaron 2006). Crump et al. (2002) claimed

that animal feeds were a source of contamination and could lead to transmission of

Salmonella to humans. Although there are over 2,500 serovars of Salmonella, there are

very few pathogenic strains which may be found across rendered feed ingredients, farm

animals and humans. Knox et al. (1963) established a connection between a Salmonella

Heidelberg outbreak from contaminated milk and the meat and bone meal used in the

feed supplied to the milk-producing cattle. In 2010, an egg recall due to Salmonella

Enteriditis contamination was initially blamed on rendering materials by the farmer

implicated but a thorough investigation proved that rendering products were not the

source for this outbreak (Caparella 2010).

In 2010, FDA identified eight Salmonella serotypes as pathogenic to animals and

listed those serotypes as of concern for potential transmission through animal feeds

(FDA, 2010). The organisms of concern associated with poultry are Salmonella

Pullorum, Salmonella Gallinarum, and Salmonella Enteritidis. The organism(s) of

concern for swine is Salmonella Choleraesuis, for sheep is Salmonella Abortusovis, for

11

horses is Salmonella Abortusequi and for cattle are Salmonella Newport and Salmonella

Dublin (FDA 2010).

The rendering industry created the Animal Protein Producers Industry (APPI) in

1984 to promote biosecurity in rendered animal feeds and reduce incidence of Salmonella

(Meeker and Hamilton, 2006). In 2004, the APPI Code of Practice certification program

for rendering plants was developed and currently more than 100 rendering plants are

certified. The APPI Code of Practice Seminar is an educational series of training courses

which teaches rendering plant workers handling and processing procedures to produce

safe feed ingredients (Meeker and Hamilton, 2006).

In 1993 and 1994, FDA conducted two separate studies to examine rendered

animal feed products for the presence of Salmonella enterica and determined 56% and

25% of the samples, respectively, were positive (McChesney et al. 1995; Crump et al.

2002). Troutt et al. (2001) examined 17 rendering facilities located in seven midwestern

states of the United States. This study also reported that a majority of raw tissue samples

entering rendering facilities were positive for Clostridium, Listeria, and Salmonella

species. No Salmonella was found in crax samples or in the rendering processing

environment. The finished rendered products contained 12 serovars of Salmonella.

Franco (2005) analyzed approximately 200 rendered animal protein meal samples over a

12 mo period for the presence of Salmonella species, and reported that Salmonella cells

were present in low numbers in animal feed at a median level of 0.09 MPN/g. Kinley et

al. (2009) examined products from 12 rendering facilities in the United States and

detected 13 Salmonella serotypes. Kinley et al. (2010) conducted a research survey to

12

determine the prevalence of Salmonella and Enterococcus species in rendering products

from 12 rendering companies. Enterococcus species were detected in 81.3% of the

samples. Salmonella was detected in 8.7% of the samples. However, 13 serotypes of

Salmonella including Senftenberg, Oranienburg, Idikan, Johannesburg, IIIa. 42:z4,z23,

Banana, Demerara, Putten, Molade, Montevideo, Mbandaka, Livingstone, and

Amsterdam were characterized by 16 pulsed-field gel electrophoresis patterns. Each set

of pulsed-field gel electrophoresis patterns was compared between product type and

rendering plant to demonstrate there was not one particular serotype present in a

particular rendering facility over a seven mo period. This study suggested the presence of

Salmonella in the finished products may be due to post-processing contamination. The

results from Franco (2005) and Kinley et al. (2010) indicated the efforts taken by the

rendering industry have microbiologically improved its products since the studies

conducted by FDA in the 1990s.

Contamination with Salmonella species in a rendering facility may be due to

cross-contamination from the raw animal tissue during processing (Ockerman and

Hansen 2000). Incoming raw rendering materials from animals serve as a reservoir for

many pathogenic bacterial species including Staphylococcus species, Listeria species,

Bacillus species, Clostridium species, Mycobacterium species, Enterobacteriaceae,

Pseudomonas species, Aeromonas species, Plesiomonas shigelloides, and Vibrio species

which can survive and exist in animals, particularly in their digestive tracts (Jay 2005).

Depending on carcass size, raw materials may be ground to reduce particle size prior to

the cooking process. Aerosols generated during the grinding process have the potential to

13

spread contaminating bacteria such as Salmonella throughout the rendering plant,

including areas where the finished product is handled (ICMSF 2000; Swingler 1982).

Jones and Bradshaw (1996) observed the strain Salmonella Enteriditis and its capability

of producing biofilms on environmental surfaces which could serve as a reservoir for

future contamination.

Emulsifiers

Emulsifiers are chemical additives that prevent the separation of two immiscible

liquids such as oil and water. Emulsifiers consist of molecules which have hydrophilic or

hydrophobic and lipophilic or lipophobic portions. Lecithin consists primarily of

phospholipids such as phosphatidylcholine, phosphatidylethanolamine, phosphatidyl-

inositol and phosphatidic acid in oil (Bueschelberger 2004). According to the FDA

Guidance for Industry (2006), lecithins are considered generally recognized as safe

(GRAS). They are a class of chemicals that are present in both plant and animal tissue.

The main sources of lecithins are soybean and sunflower oils (Szuhaj and List 1985;

Bueschelberger 2004). Weete et al. (1994) demonstrated lecithin had improved water/oil

emulsification after preheating to 180ºC for 90 min and subsequently mixed with a 60ºC

pre-heated water/oil phase. Zhang (2011) observed the effect of various levels of lecithin

used to emulsify high fat rendering samples in an aqueous buffer to assist in accurate

serial dilution of bacterial populations as well as the impact on the bacteria Geobacillus

stearothermophilus. Zhang (2011) determined the use of lecithin as an emulsifier in

dilution buffers appeared to be a promising method to enumerate high fat samples with

Geobacillus stearothermophilus.

14

Conclusion

On January 4, 2011, the Food Safety Modernization Act was signed into law by

President Barack Obama. This act expanded the power of the FDA to include regulation

of any aspect of food production in order to provide a safe food supply (FDA 2013). In

July 2013, the FDA released a compliance guide to inform the rendering industry of

current recommendations for pet and animal feed products. Contaminated feed products

have the potential to serve as a vehicle which can introduce pathogenic bacteria into the

food chain. Currently, the FDA can enforce regulatory actions if pet foods products are

contaminated with any serotype of Salmonella. Animal feeds contaminated with specific

infectious Salmonella serotypes can be seized and detained by the FDA. However, animal

feed contaminated with non-infectious serotypes of Salmonella will be evaluated on a

case by case basis by the FDA (FDA 2013). Therefore, conclusive data regarding the

validation of thermal lethality of rendering processes is vital to the livestock and pet food

industry and to the FDA to ensure thermal destruction of bacterial pathogens in products.

A disease outbreak in the animal livestock industry could have serious negative

consequences to the rendering industry, to the entire food animal chain, to consumers of

animal products, and to pets and their owners.

The specific objectives of this study are to 1) validate methodology for

enumerating Salmonella in high fat matrices and 2) determine the minimum thermal

requirements needed to destroy four pathogenic Salmonella serotypes (Salmonella

Choleraesuis, Salmonella Enteritidis, Salmonella Newport, and Salmonella Dublin) in

typical rendering material matrices.

15

References

Anderson, D. 2006. Rendering operations, In Essential rendering all about the animal

by-Products industry. Edited by D.L. Meeker. National Renderers

Association, Arlington, V.A. pp. 31-52.

Andrews, W.H., Bruce, V.R., June, G.A., Sherrod, P., Hammack, T.S., and

Amaguana, R.M. 2011. Chapter 5 Salmonella [online]. In FDA bacteriological

analytical manual (BAM). AOAC International, Gaithersburg, M.D. Available

from

http://www.fda.gov/Food/FoodScienceResearch/LaboratoryMethods/ucm0701

49.htm [accessed 9 April 2013].

Blackburn, C.W., Curtis, L.M., Humpheson, L., Billon, C., and McClure, P.J. 1997.

Development of thermal inactivation models for Salmonella enteritidis and

Escherichia coli 0157:H7 with temperature, pH and NaCl as controlling

factors. Int. J. Food Microbiol. 38: 31-44.

Bueschelberger, H.G. 2004. Lecithins. In Emulsifiers in food technology. Edited by

R.J. Whitehurst. Blackwell Publishing Ltd., Northampton, England, U.K. pp.

1-29.

Caparella, T. 2010. Rendered proteins not to blame for Salmonella outbreak in eggs.

Render. 2010. Available from http://www.rendermagazine.com/articles/2010-

issues/2010-october/2010-10- [accessed 9 April 2013].

Clemen, R.A. 1927. By-products in the packing industry. University of Chicago

Press, Chicago, I.L.

Crump, J.A., Griffin, P.M., and Angulo, F.J. 2002. Bacterial contamination of animal

feed and its relationship to human foodborne illness. Clin. Infect. Dis. 35:

859-865.

D’Aoust, J.Y., Aris, B.J., Thisdale, P., Durante, A., Brisson, N., Dragon, D.,

Lachapelle, G., Johnston, M., and Laidley, R. 1975. Salmonella eastbourne

outbreak associated with chocolate. Can. Inst. Food Sci. Technol. 8: 181-184.

D’Aoust, J.Y., and Purvis, U. 2009. Isolation and identification of Salmonella from

foods and environmental samples. In Compendium of analytical methods. Vol

2. Government of Canada, Ottawa, Ontario, Canada. Available from

http://www.hc-sc.gc.ca/fn-an/res-rech/analy-

meth/microbio/volume2/mfhpb20-01-eng.php [accessed 9 April 2013].

16

Dos Santos, C.I., Azevêdo, J.A., de Almeida, F.M., Pereira, L.G., de Araújo, G.G.,

Dos Santos Cruz, C.L., Nogueira, A.S., Souza, L.L., de Oliveira, G.A.

Performance and characteristics of carcass and non-carcass components of

lambs fed peach-palm by-product [online]. Trop. Anim. Health Prod.

Available from http://link.springer.com/article/10.1007%2Fs11250-013-0424-

1/fulltext.html [accessed 26 September 2013].

Droffner, M.L., and Yamamoto, N. 1992. Procedure for isolation of Escherichia,

Salmonella, and Pseudomonas mutants capable of growth at the refractory

temperature of 54ºC. J. Microbiol. Methods. 14: 201-206.

Franco, D.A. 1997. The genus Salmonella. Sanitation and hygiene in the production

of rendered animal by products. Animal Protein Producers Industry, National

Renderers Association, Fats and Proteins Research Foundation, Alexandria,

V.A.

Franco, D.A. 2005. A survey of Salmonella serovars and most probable numbers in

rendered-animal-protein meals: inferences for animal and human health. J.

Environ. Health. 67: 18-23.

Glenn, L.M. 2006. Isolation and identification of thermally resistant bacteria in raw

poultry rendering materials. M.S. Thesis, Clemson University. Clemson. S.C.

Goepfert, J.M., Iskander, I.K., and Amundson, C.H. 1970. Relations of the heat

resistance of Salmonellae to the water activity of environment. Appl.

Microbiol. 19: 429-433.

Heldman, D., and Hartel, R. W. (Editors). 1998. Principles of food processing. Aspen

Publishers, Gaithersburg, M.D.

Humphrey, T.J., Richardson, N.P., Statton, K.M., and Rowbury, R.J. 1993. Effects of

temperature shift on acid and heat tolerance in Salmonella enteritidis phage

type 4. Appl. Environ. Microbiol. 59: 3120-3122.

International Commission on Microbiological Specifications for Foods (ICMSF).

1986. Microorganisms in foods 6. University of Toronto Press, Toronto,

Canada.

Jay, J.M. 2005. Modern food microbiology, 6th Edition. Aspen Publishers Inc.,

Gaithersburg, M.D.

Jones, K., and Bradshaw, S.B. 1996. Biofilm formation by the Enterobacteriaceae: a

comparison between Salmonella enteritidis, E. coli and a nitrogen fixing strain

of Klebsiella pneumoniae. J. Appl. Bacteriol. 80: 458-464.

17

Juneja, V.K., Eblen, B.S., and Ransom, G. 2000. Thermal inactivation of Salmonella

spp. in chicken broth, beef, pork, turkey and chicken: determination of D- and

Z-values. J. Food Sci. 66: 146-152.

Kinley, B. 2009. Prevalence and biological control of Salmonella contamination in

rendering plant environments and the finished rendered meals. Ph.D.

Dissertation. Clemson University, Clemson, S.C.

Kinley, B., Rieck, J., Dawson, P., and Jiang, X. 2010. Analysis of Salmonella and

enterococci isolated from rendered animal products. Can. J. Microbiol. 56: 65-

73.

Kirby, R.M., and Davies, R. 1990. Survival of dehydrated cells of Salmonella

typhimurium LT2 at high temperatures. J. Appl. Microbiol. 68: 241-246.

Knox, W.A., Galbraith, N.C., Lewis, M.J., Hickie, G.C., and Johnston, H.H. 1963. A

milkborne outbreak of food poisoning due to Salmonella Heidelberg. J. Hyg.

61: 175-185.

Le Minor, L., Coynault, C., Rhode, R., Rowe, B., and Aleksic, S. 1973. Localisation

plasmidique de determinant genetique du caractere atypique “saccharose +”

des Salmonella. Ann. Microbiol. 124B: 295-306.

Le Minor, L., Coynault, C., and Pessoa, G. 1974. Determinisme plasmidique du

caractere atypique “lactose postif” de souches de S. typhimurium et de S.

oranienburg isolates au Bresil lors d’epidemies de 1971 a 1973. Ann.

Microbiol. 125A: 261-285.

Lowry, P.D., Fernando, T., and Gill, C.O. 1979. Spore survival during batch dry

rendering of abattoir waste. Appl. Environ. Microbiol. 38: 335-336.

Ma L., Zhang, G., Gerner-Smidt, P., Mantripragada, V., Ezeoke, I., and Doyle, M.P.

2009. Thermal inactivation of Salmonella in peanut butter. J. Food Prot. 72:

1596–1601.

Mackey, B.M., and Derrick C.M. 1986. Elevation of the heat resistance of Salmonella

typhimurium by sublethal heat shock. J. Appl. Bacteriol. 61: 389-393.

Mackey, B.M., and Derrick, C. 1990. Heat shock protein synthesis and

thermotolerance in Salmonella typhimurium. J. Appl. Bacteriol. 69: 373-383.

McChesney, D.G., Kaplan, G., and Gardner, P. 1995. FDA survey determines

Salmonella contamination. Feedstuffs. 67: 20-23.

18

Meeker, D.L., and C.R. Hamilton. 2006. An overview of the rendering industry. In

Essential rendering all about the animal by-products industry. Edited by D.L.

Meeker. National Renderers Association, Arlington, V.A. pp. 1-16.

Mossel, D.A.A., Jongerius, E., and Koopman, M.J. 1965. Sur la nécessité d'une

revivification préalable pour le dénombrement des Enterobacteriaceae dans les

aliments déshydratés, irradiés ou non. Ann. Inst. Pasteur Lille. 16: 119–125.

Murphy, R.Y., Duncan, L.K., Johnson, E.R., Davis, M.D., and Smith, J.N. 2002.

Thermal inactivation D- and Z-Values of Salmonella serotypes and Listeria

innocua in chicken patties, chicken tenders, franks, beef patties, and blended

beef and turkey patties. J. Food Prot. 65: 53–60.

Murphy, R.Y., Osaili, T., Duncan, L.K., and Marcy, J.A. 2004. Thermal inactivation

of Salmonella and Listeria monocytogenes in ground chicken thigh/leg meat

and skin. Poult. Sci. 83: 1218-1225.

Ng, H., Bayne, H. G., and Garibaldi, J. A. 1969. Heat resistance of Salmonella: the

uniqueness of Salmonella senftenberg 775W. Appl. Environ. Microbiol. 17:

78-82.

Ockerman, H.W., and Hansen, C.L. 2000. Animal by-product processing and

utilization. CRC Press, Taylor & Francis Group, Boca Raton, F.L.

Orta-Ramirez, A., Price, J.F., Hsu, Y. Veeramuthu, G.J., Cherry-Merritt, J.S. and

Smith, D.M. 1997. Thermal inactivation of Escherichia coli O157:H7,

Salmonella senftenberg, and enzymes with potential as time-temperature

indicators in ground beef. J. Food Prot. 60: 471-475.

Park, E.J., and Kang, D.H. 2008. Fate of Salmonella Tennessee in peanut butter at 4

and 22ºC. J. Food Sci. 73: 82-86.

Rasmussen, O.G., Hansen, R., Jacobs, N.J., and Wilder, O.H.M.1964. Dry heat

resistance of Salmonellae in rendered animal by-products. Poult. Sci. 43:

1151-1157.

Reinman, H. 1968. Effect of water activity on the heat resistance of Salmonella in

"dry" materials. Appl. Microbiolol. 16: 1621-1622.

Richardson, L. 2006. Convert waste into biofuel. California Farmer. 2006. Available

from http://magissues.farmprogress.com/CLF/CF02Feb06/clf26.pdf [accessed

on 9April 2013].

19

Swingler, G.B. 1982. Meat microbiology. Edited by M.H. Brown. Applied Science

Pub., London, England, U.K.

Szuhaj, B.F., and List, G.R. 1985. Lecithins. American Oil Chemists’ Society,

Urbana, I.L.

Teixeira. A.A. 2006. Thermal food processing. Edited by D. Sun. CRC Press, Taylor

& Francis Group, Boca Raton, F.L.

Troutt, H.F., Schaeffer, D., Kakoma, I., and Pearl, G.G. 2001. Prevalence of selected

foodborne pathogens in final rendered products. FPRF Directors Digest #312,

Fats and Proteins Research Foundation, Bloomington, I.L.

United States Food and Drug Administration (FDA). 2006. Guidance for industry:

guidance on the labeling of certain uses of lecithin derived from soy under

section 403(w) of the federal food, drug, and cosmetic act. United States FDA.

Rockville, M.D. Available from

http://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulator

yInformation/LabelingNutrition/ucm059065.htm [accessed 9 April 2013].

United States Food and Drug Administration (FDA). 2010. Compliance policy guide

sec. 690.800 Salmonella in animal feed. Draft guidance. United States FDA.

Rockville, M.D. Available from http://www.gpo.gov/fdsys/pkg/FR-2010-08-

02/pdf/2010-18873.pdf [accessed on 9 May 2013].

United States Food and Drug Administration (FDA). Bad bug book.2012. Food-borne

pathogenic microorganisms and natural toxins handbook. United States

Department of Human and Health Services. United States FDA. Rockville,

M.D. Available from

http://www.fda.gov/downloads/Food/FoodborneIllnessContaminants/UCM29

7627.pdf [accessed 9 May 2013].

United States Food and Drug Administration (FDA). 2013. Compliance policy guide

sec. 690.800 Salmonella in food for animals. United States FDA. Rockville,

M.D. Available from

https://www.federalregister.gov/articles/2013/07/16/2013-16975/compliance-

policy-guide-sec-690800-salmonella [accessed on 9 August 2013].

Weete, J.D., Betageri, S., and Griffith, G.L. 1994. Improvement of lecithin as an

emulsifier for water-in-oil emulsions by thermalization. J. Am. Oil Chem.

Soc. 71: 731–737.

Wray C., and Wray, A. 2000. Salmonella in domestic animals. CABI Publishing,

Wallingford, Oxon, United Kingdom.

20

Zhang, Y. 2011. Thermal destruction of Geobacillus stearothermophilus in rendered

animal co-products. M.S. thesis. Clemson University, Clemson, S.C.

21

CHAPTER 2

METHOD VALIDATION FOR ENUMERATING BACTERIA IN HIGH FAT

MATRICES

Abstract

The high fat content of rendered animal co-products has complicated traditional

bacterial enumeration methodology. Therefore, it is vital that the rendering industry has

accurate enumeration methodologies for pathogenic bacteria in finished products. An

objective of this study was to examine the use of the standard Class O phosphate/

magnesium chloride dilution series (Dilution Series A) and a modified (pre-warmed to

32ºC) lecithin phosphate dilution buffer series (Dilution Series B) by comparing mean

bacterial counts of a Salmonella cocktail in poultry and beef rendering materials. The

results of this study did not indicate that the use of a modified buffer to improve

enumeration of Salmonella from poultry and beef rendering materials. Instead, the results

suggested that the use of xylose lysine deoxycholate agar (XLD) with either buffer

system would produce accurate enumeration data of Salmonella from poultry and beef

rendering materials.

22

Introduction

Rendering is the process of converting inedible animal tissue from food animals

into granular meals and liquid fats that are used in numerous co-products, including

animal feeds, chemical feedstocks, and fertilizers (Meeker and Hamilton 2006). The

continuous cooking process used by the rendering industry is reported to be 40 to 90 min

at 240 to 290ºF (115.6 to 143.3ºC) (Meeker and Hamilton, 2006). The high temperatures

used in the rendering cooking process reduce the number of microorganisms in raw

perishable animal tissues. Marginal processing conditions could result in the presence of

residual microorganisms in finished products (Crump et al., 2002). Crump et al. (2002)

indicated that animal feeds can be a source of contamination of Salmonella to humans.

Therefore, it is vital to develop accurate enumeration methods for high fat rendering

materials.

The high fat content of rendered animal co-products has complicated traditional

bacterial enumeration methodology, making it difficult to accurately determine the

presence or absence of Salmonella in rendering co-products (Glenn 2006). It is

hypothesized that this fat content could also entrap the bacterial cells in rendering

materials. Therefore, upon serial dilution, the fat globules may not be evenly dispersed

throughout dilutions and subsequently not be transferred evenly to plates for

enumeration. Inaccurate transference to the microbial media would yield either higher or

lower bacterial counts and overall a less accurate method of enumeration. Zhang (2011)

determined that the use of lecithin as an emulsifier in dilution buffers appeared to be a

promising method to enumerate high fat samples with Geobacillus stearothermophilus.

23

Lecithin consists of complex combinations of phospholipids and is a common emulsifier

in the food industry (Bueschelberger 2004).

An objective of the study was to examine the use of the standard Class O

phosphate/magnesium chloride dilution buffer series and a modified (pre-warmed to

32ºC) lecithin phosphate dilution buffer series by comparing mean bacterial counts of a

Salmonella cocktail in each poultry and beef rendering materials, adjusted to 50% fat

content.

Materials and Methods

Rendering Sample Preparation

Samples of poultry and beef rendering fat and crax materials were collected on

three separate days from rendering plants in the midwestern and southeastern U.S. Crax

is a solid material composed of protein, minerals, and residual fat that is discharged from

the screw press during the rendering process and is typically further ground into meat and

bone meal (Meeker and Hamilton, 2006). Crax samples were submitted in duplicate to

the Clemson University Agricultural Services Laboratory for ash, fat, and moisture

content analysis. The crax and fat samples were re-mixed to produce 50% fat samples. A

food processor was disinfected by rinsing in Antibac B™ (Diversey Corporation,

Cincinnatti, OH) dissolved in distilled deionized water (ddH2O) (0.6 g per L) for

approximately 2 min, followed by rinsing 5 times with sterile ddH2O. Particle size was

reduced by processing for approximately 10 min on the pulse setting in a disinfected food

processor (Robot Coupe Model R2 Ultra, Ridgeland, MS) prior to conducting the

experiments. A sterile stainless steel spatula was used to scrape material from the sides

24

during pauses in processing. All samples were stored under refrigeration until needed for

experimentation.

Salmonella Preparation

Four pathogenic Salmonella recognized by FDA as hazardous for animal feeds

(Salmonella Choleraesuis (FDA 8326) (SC), Salmonella Enteritidis (USDA H4386) (SE),

Salmonella Newport (USDA H1073) (SN) and Salmonella Dublin (FDA 23742) (SD))

were obtained for this study (FDA 2010; FDA 2013). SE and SN were obtained from Dr.

Vijay Jejuna of the USDA Agricultural Research Service, Microbial Food Safety

Research Unit, 600 East Mermaid Lane, Room 2129, Wyndmoor, PA 19038. SC and SD

were obtained from the food microbiology culture collection of collaborator Dr. Xiuping

Jiang at Clemson University.

A preliminary study was conducted to determine the optimal media conditions for

Salmonella growth. Trypticase soy broth (TSB) (90000-050, VWR Scientific Products,

Suwanee, GA), TSB with the addition of 0.1% (wt/vol) yeast extract (MP Biomedicals,

LLC, Solon, Ohio), and brain heart infusion broth (BHI) (211059, VWR Scientific

Products, Suwanee, GA) were tested. TSB with the addition of 0.1% (wt/vol) yeast

extract was chosen as the best media. The media choice was based on highest cell

densities determined from optical density measurements (µQuant Universal Microplate

Spectrophotometer, Bio-Tek Instruments, Winooski, VT) at 600 nm and dilution plating

in duplicate onto bismuth sulfite agar (90003-904, VWR Scientific Products, Suwanee,

GA), Hektoen enteric agar (9004-054, VWR Scientific Products), xylose lysine

deoxycholate (XLD) (90003-996, VWR Scientific Products), and trypticase soy agar

25

(TSA) (90000-050, VWR Scientific Products). Bismuth sulfite agar, Hektoen enteric agar

and XLD are selective media used for the detection of Salmonella in food products

(Andrews et al. 2011).

An additional preliminary study was conducted to determine if any combination

of SC, SE, SN or SD promoted or inhibited growth. Overnight cultures were adjusted to

0.5 OD at 600 nm. Flasks of sterile TSB with 0.1% yeast extract were inoculated with

equal volumes of each Salmonella serotype or combinations of the four serotypes.

Cultures were incubated overnight at 35ºC and the OD was measured again to determine

if growth had increased or stayed the same. The results indicated that no combination of

SC, SE, SN, and SD appeared to enhance or inhibit growth.

For the study, each serotype was grown individually in 1 L TSB (90000-050,

VWR Scientific Products, Suwanee, GA). Each overnight culture was washed twice by

centrifugation at 7,000 x g for 7 min (GSA rotor, DuPont RC5C Sorvall Instruments

Centrifuge, DuPont Company, Newtown, CT) and resuspended in sterile physiological

(0.85%) saline. Optical density was adjusted to 0.7 (ca. 108 cfu/mL) at 600 nm and equal

volumes of the four cultures were combined in a sterile flask.

Salmonella Enumeration in Rendering Materials

The standard Class O phosphate/magnesium chloride (Wehr and Frank 2004)

dilution buffer system (Dilution Series A using diluent a) was compared to a modified

dilution system (Dilution Series B) (Fig. 2.1). Dilution Series B was comprised of two

modified phosphate/magnesium chloride dilution buffers containing lecithin (AA36486-

A1, Alfa Aesar, Ward Hill, MA) at the rate of 2 g per 99 mL (diluent b) and 0.5 g per 99

26

mL (diluent c), respectively, for dilution of high fat materials followed by use of standard

Class O phosphate/magnesium chloride buffer (diluent a) (Fig. 2.1). Dilution Series A

and Dilution Series B were prepared, dispensed in 99 mL aliquots into dilution bottles,

and autoclaved (Fig. 2.1). Prior to experimentation, Dilution Series A was stored and

used at room temperature (Fig. 2.1). Dilution Series B was pre-warmed to 32ºC (Fig.

2.1).

A preliminary experiment was conducted to validate the use of 1 mL of a

Salmonella cocktail in 20 g of each poultry and beef rendering (50% fat) sample. One mL

of crystal violet dye (90008-894, VWR Scientific Products, Suwanee, GA) was added to

20 g of each poultry and beef rendering (50% fat) sample in a sterile Whirl-Pak™ sample

bag (11216-409, VWR Scientific Products, Suwanee, GA). Crystal violet dye was used to

represent the culture. The mixtures were stomached (Laboratory Blender, Stomacher 400,

A.J. Seward and Co. Ltd., London, England) for 2 min on the high setting. Subsamples

were observed for color uniformity using a microscope (Carl Zeiss, Photomicroscope III,

Oberkochen, West Germany) at 10x and 40x magnification. Results indicated that a 1:20

ratio of culture to sample would allow for even distribution of culture throughout each of

the poultry and beef rendering samples.

One mL (ca. 108

cfu./mL) of a Salmonella cocktail was added to each 20 g poultry

and beef rendering sample (50% fat content) in a sterile Whirl-Pak™ sample bag. The

mixtures were stomached for 2 min on the high setting. Subsamples of the mixture were

diluted using each Dilution Series A and Dilution Series B. Dilutions were carried out to

the 10-9

dilution (Fig. 2.1) and plated in duplicate onto bismuth sulfite agar, Hektoen

27

enteric agar, XLD, and TSA. Controls included media and diluent sterility controls as

well as uninoculated rendering samples (50% fat). Plates were incubated overnight at

35ºC and enumerated.

Statistical Analysis

The mean bacterial counts of the Salmonella cocktail obtained from the culture

controls and inoculated samples were converted to log10 cfu/g values ± standard error.

The mean bacterial counts of the culture controls diluted in the standard Dilution Series A

and the mean bacterial counts of the inoculated samples plated onto the same media were

compared using a two-tailed, paired Student's t tests in Microsoft Excel®

(Microsoft®

,

2010) to determine statistical significance at alpha=0.05.

Results

The analysis of the beef rendering materials (n=6) indicated the average fat

content ranged from 9.9% to 13.8%, average ash content ranged from 20.6% to 33.5%,

and average moisture content ranged from 2.1% to 3.3%. Averaged analysis data for each

pair of duplicate samples (Day 1, Day 2, Day 3) were used to prepare 50% fat materials for

this study.

Despite the type of media, the mean bacterial counts obtained from Salmonella

cocktail culture controls diluted in Dilution Series A were not significantly different

(P<0.05) from the mean bacterial counts of the Salmonella cocktail culture controls

diluted in the standard Dilution Series B (Table 2.1). The mean bacterial counts of

6.01±0.28 and 5.77±0.30 log10 cfu/g were obtained from the inoculated poultry rendering

samples diluted with each Dilution Series A and Dilution Series B, respectively, and

28

plated onto bismuth sulfite agar. However, the mean bacterial counts in poultry rendering

materials were statistically different than the Salmonella culture control when it was

diluted with the standard Dilution Series A and plated onto bismuth sulfite agar (P<0.05)

(Table 2.1).

Enumeration data revealed mean bacterial counts of 7.46±0.99 and 6.48±1.00

log10 cfu/g from the inoculated poultry rendering samples diluted with each Dilution

Series A and Dilution Series B, accordingly, and plated onto Hektoen Enteric agar. These

mean bacterial counts were not statistically different than the Salmonella culture control

diluted with the standard Dilution Series A and plated onto Hektoen Enteric agar

(P<0.05) (Table 2.1).

The mean bacterial counts of 8.14±1.76 and 7.81±1.45 log10 cfu/g obtained from

the inoculated poultry rendering samples diluted with each Dilution Series A and Dilution

Series B, respectively, and plated onto XLD. These mean bacterial counts were not

statistically different than the Salmonella culture control diluted with the standard

Dilution Series A and plated onto XLD (P<0.05) (Table 2.1).

Enumeration of inoculated poultry rendering samples diluted with each Dilution

Series A and Dilution Series B revealed the mean bacterial counts of 6.80±0.88 and

7.10±0.85 log10 cfu/g, accordingly, on TSA. These mean bacterial counts were not

statistically different than the Salmonella culture control diluted with the standard

Dilution Series A and plated onto TSA (P<0.05) (Table 2.1).

The mean bacterial counts of 7.45±0.99 and 6.47 ±0.99 log10 cfu/g obtained from

the inoculated beef rendering samples diluted with each Dilution Series A and Dilution

29

Series B, respectively, were not statistically different than the Salmonella culture control

on bismuth sulfite agar (P<0.05) (Table 2.1).

Enumeration data revealed the mean bacterial counts of 5.53±0.03 and 5.53 ±0.03

log10 cfu/g from the inoculated beef rendering samples diluted with each Dilution Series

A and Dilution Series B, accordingly, and plated on Hektoen Enteric agar. These mean

bacterial counts were statistically different than the Salmonella culture control on

Hektoen Enteric agar (P<0.05) (Table 2.1).

Enumeration of inoculated beef rendering samples diluted with each Dilution

Series A and Dilution Series B revealed the mean bacterial counts of 7.14±1.67 and

7.14±1.67 log10 cfu/g, respectively, on XLD. These mean bacterial counts were not

statistically different than the Salmonella culture control on XLD (P<0.05) (Table 2.1).

The mean bacterial counts of 5.55±0.06 and 5.51±0.03 log10 cfu/g enumerated

from the inoculated beef rendering samples diluted with each Dilution Series A and

Dilution Series B, accordingly, on TSA. These mean bacterial counts were statistically

different than the Salmonella culture control on TSA (P<0.05) (Table 2.1).

Discussion

In this study, Dilution Series A and Dilution Series B were used to enumerate the

Salmonella cocktail. The mean bacterial counts obtained from the culture controls

enumerated with each dilution series were not statistically different despite the media

used (P<0.05) (Table 2.1). There results indicated that the addition of the emulsifier

lecithin to the dilution buffer did not inhibit or promote the growth of the Salmonella

cocktail. The Salmonella cocktail enumeration data were compared for Dilution Series A

30

and Dilution Series B in each poultry and beef rendering materials. The mean bacterial

counts enumerated from the inoculated poultry rendering samples, serially diluted in each

Dilution Series A and Dilution Series B and plated onto bismuth sulfite agar were

significantly lower than the mean bacterial counts obtained from the Salmonella cocktail

culture on bismuth sulfite agar (P<0.05) (Table 2.1). An explanation for the lower

bacterial counts from the poultry rendering samples is not known, but could include

dilution error, presence of free fatty acids, entrapment of the bacteria in bones particles or

coating of the bacteria by fat. In previous studies on raw poultry rendering materials,

Glenn (2006) determined standard phosphate buffer serial dilutions produced irregular

microbial enumeration results. Due to the high fat content of the rendering materials, it

was revealed that the immiscibility of fat in the aqueous buffer caused the erroneous

results. Glenn (2006) indicated that the fat may have entrapped the bacteria in the

rendering materials. Additionally, the fat globules may not have dispersed evenly

throughout dilutions due to the use of aqueous buffers and subsequently not transferred

accurately to plates for enumeration (Glenn 2006). The 50% fat content in the poultry

rendering materials used in this study may have entrapped the bacterial cells leading to

lower bacterial counts.

The mean bacterial counts enumerated from the inoculated beef samples diluted

with each Dilution Series A and Dilution Series B were significantly lower (P<0.05) than

Salmonella cocktail controls on Hektoen Enteric agar and TSA (Table 2.1). Possible

reasons for the lower bacterial counts from the beef rendering samples would be the same

as above. Despite the dilution series used, these results suggested that the Salmonella

31

cocktail was accurately enumerated from beef and poultry rendering materials containing

50% fat content on XLD agar. XLD is a selective media used for the detection of

Salmonella. In this study, XLD appeared to be a better selective media for the SC, SE,

SN and SD than Hektoen Enteric or bismuth sulfite. It should also be noted that the black

Salmonella colonies on the red XLD were easier to distinguish than the black colonies on

the light green-yellow bismuth sulfite agar or the green colonies with black centers on the

dark green Hektoen enteric agar.

The results of this study did not suggest the use of a modified buffer to improve

enumeration of Salmonella from poultry and beef rendering materials. Instead, the results

suggested that the use of XLD with either buffer system would produce accurate

enumeration data of Salmonella from poultry and beef rendering materials. This research

was a preliminary step toward improving enumeration methods for the detection of

pathogenic bacterial species in high fat products.

32

Dilution Series A: Standard phosphate magnesium chloride buffer system (all bottles at

room temperature).

Dilution Series B: Modified lecithin buffer system (all bottles pre-warmed to 32ºC)

Figure 2.1. Diagram of two buffer systems used for serially diluting rendering materials.

Diluent a represents a 99 mL of phosphate/ magnesium chloride buffer. Diluent b

represents a 2 g lecithin/99 mL of phosphate/ magnesium chloride buffer. Diluent c

represents a 0.5 g lecithin/100 mL of phosphate/ magnesium chloride buffer.

1 mL 1 mL

1 mL 1 mL

1 mL 1 mL

1 g of

Sample

a a a a a

1 mL

1 g of

Sample

c a a a b

1 mL

33

Table 2.1. Comparison of mean bacterial counts from each poultry and beef rendering

material using each dilution series to a standard culture control (n=6).

Sample Dilution Series Media Mean Bacterial Count1

log10 cfu/g ± standard error

P-Value2

Salmonella A Bismuth Sulfite 7.71±0.10 aefmn

-

Salmonella A Hektoen Enteric 7.69±0.49 bghop

-

Salmonella A XLD 8.85±0.89 cijqr

-

Salmonella A TSA 8.21±0.27 dklst

-

Salmonella B Bismuth Sulfite 7.45±0.18 a 0.45

Salmonella B Hektoen Enteric 7.10±0.31 b 0.54

Salmonella B XLD 9.19±0.64 c 0.45

Salmonella B TSA 7.91±0.35 d 0.33

Poultry A Bismuth Sulfite 6.01±0.28 e 0.02

*

Poultry B Bismuth Sulfite 5.77±0.30 f 0.03

*

Poultry A Hektoen Enteric 7.46±0.99 g 0.81

Poultry B Hektoen Enteric 6.48±1.00 h 0.14

Poultry A XLD 8.14±1.76 i 0.50

Poultry B XLD 7.81±1.45 j 0.20

Poultry A TSA 6.80±0.88 k 0.15

Poultry B TSA 7.10±0.85 l 0.28

Beef A Bismuth Sulfite 7.45±0.99 m

0.84

Beef B Bismuth Sulfite 6.47±0.99 n 0.30

Beef A Hektoen Enteric 5.53±0.03 o 0.05

*

Beef B Hektoen Enteric 5.53±0.03 p 0.05

*

Beef A XLD 7.14±1.67 q 0.21

Beef B XLD 7.14±1.67 r 0.21

Beef A TSA 5.55±0.06 s 0.01

*

Beef B TSA 5.51±0.03 t 0.01

*

1 Values with the same superscripts (a-t) indicate the mean bacterial counts compared

using two-tailed, paired Student's t tests.

2indicates statistical difference at P<0.05.

34

References

Andrews, W.H., Bruce, V.R., June, G.A., Sherrod, P., Hammack, T.S., and Amaguana,

R.M. 2011. Chapter 5 Salmonella. In FDA bacteriological analytical manual

(BAM). AOAC International, Gaithersburg, M.D. Available from

http://www.fda.gov/Food/FoodScienceResearch/LaboratoryMethods/ucm070149.

htm [accessed 9 April 2013].

Bueschelberger, H.G. 2004. Lecithins. In Emulsifiers in food technology. Edited by R.J.

Whitehurst. Blackwell Publishing Ltd., Northampton, UK. pp. 1-29.

Crump, J.A., P.M. Griffin, and F.J. Angulo. 2002. Bacterial contamination of animal feed

and its relationship to human foodborne illness. Clin. Infect. Dis. 35: 859-865.


poultry rendering materials. M.S. Thesis, Clemson University. Clemson. S.C.

Meeker, D.L., and Hamilton, C.R. 2006. An overview of the rendering industry. In

Essential rendering all about the animal by-products industry. Edited by D.L.

Meeker. National Renderers Association, Arlington, VA. pp.1-17.

United States Food and Drug Administration (FDA). 2010. Compliance policy guide sec.

690.800 Salmonella in animal feed. Draft guidance. United States FDA.

Rockville, M.D. Available from http://www.gpo.gov/fdsys/pkg/FR-2010-08-

02/pdf/2010-18873.pdf [accessed on 9 May 2013].


690.800 Salmonella in food for animals. United States FDA. Rockville, M.D.

Available from https://www.federalregister.gov/articles/2013/07/16/2013-

16975/compliance-policy-guide-sec-690800-salmonella [accessed on 9 August

2013].

Zhang, Y. 2011. Thermal destruction of Geobacillus stearothermophilus in rendered

animal co-products. M.S. thesis. Clemson University, Clemson, S.C.

35

CHAPTER 3

VALIDATION OF THERMAL DESTRUCTION OF SALMONELLA IN RENDERED

BEEF PRODUCTS

Abstract

Animal rendering is a process that converts inedible animal tissue into stable,

value-added materials. The North American rendering industry annually recycles over 61

billion pounds of residual animal by-products. Approximately 85% of rendered products

are used as animal feed ingredients. Therefore, it is vital that the rendering industry has

conclusive validation data on the thermal lethality of rendering thermal processing to

destroy animal disease pathogens in finished products. The high fat, bone and protein

content of rendering materials leaves the industry with no comparable thermal death time

values from the human food industry or any other industry. The objective of this study

was to determine thermal death time values for beef rendering materials containing 50%

fat content for four pathogenic Salmonella recognized by FDA as hazardous for animal

feeds (Salmonella Choleraesuis (SC), Salmonella Enteritidis (SE), Salmonella Newport

(SN) and Salmonella Dublin (SD)). In the study, each serotype appeared to have unique

thermal death time characteristics. With increasing thermal treatment time, reduction in

the population of each serotype of Salmonella was not a straight line decrease. In fact, on

most of the cultures, after failing to detect the cultures after certain time treatments, the

culture were later detected after longer thermal treatments. In thermal treatments up to

420 s at 240ºF (115.6ºC), SC was last detected at 120 s, SE at 120 s, SN at 300 s and SD

at 360 s. However, uninoculated controls indicated thermally resistant strains in the

36

background which testing indicated were Salmonella. The presence of Salmonella or

organisms detected as Salmonella was noted up to 360 s of treatment in the uninoculated

samples. Further research will be needed to verify that these organisms are Salmonella or

some other organism that is cross-reacting. In rendering materials, bone and tissue

fragments can vary greatly across samples. In this study, a large range of particle sizes

was present in the beef rendering materials.

37

Introduction

The United States and Canadian rendering industry annually recycles over 61

billion pounds of residual animal by-products into animal feeds, fats and proteins to

prevent waste of these materials (Meeker and Hamilton, 2006). Validating thermal

lethality of rendering processes is crucial to the livestock and pet food industry and to the

FDA to ensure destruction of bacterial pathogens in products. A disease outbreak in the

animal livestock industry could have serious negative consequences to the rendering

industry and to the entire food animal chain, including consumers.

The high temperatures used in the rendering cooking process reduce the number

of microorganisms in raw perishable animal tissues. The continuous cooking process is

reported to be 40 to 90 min at 240 to 290ºF (115.6 to 143.3ºC) (Meeker and Hamilton,

2006). Crax is a solid material composed of protein, minerals, and residual fat that is

discharged from the screw press during the rendering process and is typically further

ground into meat and bone meal (Meeker and Hamilton, 2006). Meat and bone meal is

frequently used in animal feeds and pet foods. Marginal processing conditions could

result in survival of residual microorganisms in this protein rich product (Crump et al.,

2002).

Thermal death time (TDT) is a factor of time, temperature, material matrix and

organism (Heldman and Hartel, 1998). TDT is defined as the time needed to reduce a

given number of organisms at a specific temperature in a specific matrix (Jay, 2005;

Teixeira, 2006). Decimal reduction time (D value) specifies the time required for a one

log10 reduction of a particular organism at a specific temperature. The larger the D value

38

at a given temperature, the higher the thermal resistance of the microbial population

(Heldman and Hartel, 1998). The high fat, bone and protein content of rendering

materials leaves the rendering industry with no comparable thermal death time values

from the human food industry or any other industry. The objective of this study was to

determine the TDT and D values for beef rendering materials containing 50% fat content

for four pathogenic Salmonella recognized by FDA as hazardous for animal feeds

(Salmonella Choleraesuis, Salmonella Enteritidis, Salmonella Newport, and Salmonella

Dublin) (FDA, 2010; FDA, 2013) at 240ºF (115.6ºC).



Samples of beef crax and beef tallow were obtained from a midwestern rendering

company on three separate days. The crax samples were submitted in duplicate to the

Clemson University Agricultural Service Laboratory for ash, fat, and moisture content

analysis. The crax and tallow samples were re-mixed to produce 50% fat samples. A food

processor bowl, blade and lid were disinfected by rinsing in Antibac B™ (Diversey

Corporation, Cincinnatti, OH) dissolved in distilled deionized water (ddH2O) (0.6 g per

L) for approximately 2 min, followed by rinsing 5 times with sterile ddH2O. Particle size

was reduced by processing for approximately 10 min on the pulse setting in the

disinfected food processor (Robot Coupe Model R2 Ultra, Ridgeland, MS) prior to

conducting the experiments. A sterile stainless steel spatula was used to scrape material

from the sides during pauses in processing. All samples were stored under refrigeration

until needed for experimentation.

39


Four pathogenic Salmonella serotypes recognized by FDA as hazardous for

animal feeds (Salmonella Choleraesuis (FDA 8326) (SC), Salmonella Enteritidis (USDA

H4386) (SE), Salmonella Newport (USDA H1073) (SN) and Salmonella Dublin (FDA

23742) (SD)) were obtained for this study (FDA, 2010; FDA, 2013). SE and SN were

obtained from Dr. Vijay Jejuna of the USDA Agricultural Research Service, Microbial

Food Safety Research Unit, 600 East Mermaid Lane, Room 2129, Wyndmoor, PA 19038.

SC and SD were obtained from the food microbiology culture collection from

collaborator Dr. Xiuping Jiang at Clemson University.






extract was chosen as the best media based highest cell densities determined from optical

density measurements (µQuant Universal Microplate Spectrophotometer, Bio-Tek

Instruments, Winooski, VT) at 600 nm and dilution plating in duplicate onto onto

bismuth sulfite agar (90003-904, VWR Scientific Products, Suwanee, GA), Hektoen

enteric agar (9004-054, VWR Scientific Products), xylose lysine deoxycholate (XLD)

(90003-996, VWR Scientific Products), and trypticase soy agar (TSA) (90000-050, VWR

Scientific Products).

40

As a preliminary study, each individual Salmonella serotype was plated onto

bismuth sulfite agar, Hektoen enteric agar, XLD, and TSA. Enumeration data indicated

use of XLD and TSA as the preferred agar media for enumerating SC, SE, SN, and SD.

A preliminary goal of this experiment was to obtain concentrated bacterial slurry

of each serotype to use in inoculating beef rendering materials for thermal processing.

The average concentrations of Salmonella cultures in broth for SC, SE, SN, and SD after

24 h incubation at 35ºC were 8.66±0.02, 8.56±0.03, 8.80±0.06, and 8.65±0.03 log10

cfu/g, respectively. Preliminary experiments were conducted to determine the volume of

culture as well as concentration rate necessary. Enumeration on XLD and TSA verified

that 5 L of a 24 h Salmonella culture grown in TSB with 0.1% (wt/vol) yeast extract and

then concentrated by centrifugation was optimal. Centrifugation was conducted at 7,000

x g for 7 min (GSA rotor, DuPont RC5C Sorvall Instruments Centrifuge, DuPont

Company, Newtown, CT) at 4ºC in sterile centrifuge bottles (47735-696, VWR Scientific

Products, Suwanee, GA) and the supernatant was discarded after autoclaving. The pellet

was resuspended in 5 mL sterile TSB. In preliminary studies conducted 3 times in

duplicate (n=6), the average bacterial concentrations after centrifugation and

resuspension for SC, SE, SN, and SD were determined. This procedure was used to

prepare the bacterial cultures used throughout the experiment.

Each slurry of Salmonella, prepared as above, was inoculated into beef rendering

material at the rate of 100 µL culture per 1 g sample. In a preliminary study, two methods

were conducted. The mean bacterial counts of each concentrated bacterial slurry and the

inoculated samples were determined. Method 1 was the serial dilution of each bacterial

41

slurry and each inoculated sample to 10-14

utilizing the standard Class O

phosphate/magnesium chloride dilution buffer (Wehr and Frank, 2004). Method 2 was

the serial dilution of each bacterial slurry and sample to 10-14

using pre-warmed (32ºC)

modified Class O phosphate/magnesium chloride diluent. Controls included media and

uninoculated beef rendering samples. Each experiment was conducted 3 times in

duplicate (n=6).

Thermal Death Time Trials

Stainless steel sample tubes (8.5 cm length, 1.6 cm outer diameter, 1.3 cm inner

diameter) were custom manufactured by a local company by boring 304 stainless steel

rods. The tubes were capped (60825-801, VWR International, Suwanee, GA) and

autoclaved. Beef rendering samples (50% fat) were aseptically transferred (1 g) into

sixteen sterile tubes. The tubes were placed in an analog dry block heater (Model

#12621-108, VWR International, Suwanee, GA) equipped with Model #13259-162

heating blocks (VWR International, Suwanee, GA) set to 115.6ºC. Four of the tubes were

randomly selected as temperature controls using dial thermometers (61159-409, VWR

Scientific Products, Suwanee, GA). The tubes were heated to an internal treatment

temperature of 115.6ºC prior to addition of the cultures. Each individual culture (100 µL)

was directly pipetted into 1 g of the heated rendering samples. After culture inoculation,

the sample was pipetted up and down approximately four times to thoroughly mix. Upon

inoculation and mixing, time measurements (0, 15, 30, 60, 90, 120, 180, 240, and 300 s)

were started on the thermal treatment. After preliminary experiments on SN and SD

indicated longer thermal treatment was needed, additional trials were included for the

42

time treatments of 0, 90, 240, 300, 360 and 420 s for these cultures. Samples were placed

on ice immediately after thermal treatment. Additional sample tubes containing beef

rendering used for unheated controls were placed on ice until used for plating. All

samples were processed for microbial content immediately after conclusion of heat

treatments.

A preliminary experiment was conducted to validate the use of 1 g of sample pre-

enriched in 5 mL of sterile universal pre-enrichment broth (UPB) (95021-036, VWR

Scientific Products, Suwanee, GA) in comparison to 1 g of sample pre-enriched in 9 mL

of UPB as recommended by the FDA Bacteriological Analytical Manual (BAM)

(Andrews et al., 2011). The stainless steel tubes used in this experiment would not hold

the 1 g of sample pre-enriched plus 9 mL of UPB. Results indicated that the 1:5 ratio of

sample to pre-enrichment broth was as effective as the 1:9 ratio of sample to pre-

enrichment broth. Therefore, this procedure was used throughout the experiment.

Once 5 mL of sterile UPB was aseptically pipetted into each tube, the wooden

shaft of a sterile cotton-tipped applicator (89133-814, VWR Scientific Products,

Suwanee, GA) was used to thoroughly mix the sample for 30 s. Each UPB diluted sample

(0.1 mL) was directly pipetted onto XLD and TSA plates and spread using an alcohol-

flamed bent glass rod. As a control, each Salmonella slurry was serially diluted to 10-12

in

the standard Class O phosphate/magnesium chloride dilution buffer and either 1.0 mL or

0.1 mL was spread plated onto XLD and TSA. Media and dilution buffer controls also

were conducted. All plates were incubated overnight at 35ºC. In this experimental design,

XLD selected for Salmonella spp. while TSA measured total aerobic, mesophilic

43

bacterial counts. This included any background bacteria and, in the test samples,

background bacteria plus inoculated Salmonella. For each inoculated or uninoculated

beef rendering sample, dilutions were carried out such that the direct plating on XLD and

TSA had a lower detection limit of 1.4 log10 cfu/g.

Because the direct plate counting method had a lower detection limit of 1.4 log10

cfu/g, an additional experiment was conducted in accordance with the FDA BAM

procedures to detect as low as 1 cfu/g (Andrews et al., 2011). The remaining UPB diluted

sample in the stainless steel tube was incubated overnight at 35ºC and then vortexed

(Super Mixer, 1290, Labline Instruments, Inc., Melrose Park, IL) on the fast setting for

approximately 30 s. The sample was aseptically pipetted (0.1 mL) to Rappaport-

Vassiliadis (RV) pre-enrichment broth (10 mL) (95039-382, VWR Scientific Products,

Suwanee, GA). The same sample was aseptically pipetted (1 mL) to tetrathionate broth

(TT) (10 mL) (90000-008, VWR Scientific Products, Suwanee, GA). Controls included

the concentrated bacterial slurry and sterile media. The samples and control broth were

incubated overnight at 42ºC. A 3 mm inoculation loop of each pre-enriched sample and

control was streaked onto XLD. All plates were incubated overnight at 35ºC. Results

indicated the presence or absence of Salmonella in the samples. As per FDA BAM,

positive samples obtained from the RV and TT pre-enrichments were validated using two

confirmation tests (Feng, 2001). Latex agglutination tests (FT0203, Thermo Fisher

Scientific, Waltham, MA 02454) and ChromAgar™ (90006-158, VWR Scientific

Products, Suwanee, GA) were conducted using each Salmonella culture as a control (BD

Diagnostics, 2008; Oxoid Limited, 2013). In order to analyze the data, when duplicate

44

results from the pre-enriched samples were both negative the data was reported as 0.0

(Fig. 3.1). If one duplicate was positive and one was negative, it was reported as 0.5. If

both duplicates were positive, it was reported as 1.0 (Fig. 3.1).

Bone Particle Size Determination

To determine the variation in bone particle size in the processed 50% fat

rendering material used, 10 g of the rendering sample was sized through a series of sieves

(57333-965, VWR Scientific Products, Suwanee, GA) equipped with eight different

standard mesh sizes (25, 35, 45, 60, 80, 120, 170 and 230 µm). Samples were measured

into the upper sieve and processed using 100 mL of hexane (AAAL13233-AU, VWR

Scientific Products, Suwanee, GA) to dissolve fat and assist in particle separation. The

hexane fraction was washed through the sieve column 10 times. Each fraction of particle

size was reported as a percentage of the total weight of the rendering sample. Each trial

was repeated 10 times per day for 3 days (n=30).

Determination of Estimated D Values

The direct plate count of each concentrated Salmonella slurry and the time at

which each culture was destroyed were compared on graphs. In a preliminary experiment,

percent recoveries of Salmonella from inoculated beef samples were calculated for each

recoverable Salmonella population density. Due to the experimental design, the actual

population count from beef rendering material was not conducted. However, the total

count in each bacterial slurry was measured. This population count was used in estimated

D value calculations. The final time the population was no longer detected in each RV

and TT pre-enrichment as validated by the two confirmation tests was used as the thermal

45

death time. These data were graphed and the slope of the line was used to calculate the

estimated D value.

Results

Analysis of beef rendering materials indicated fat content ranged from 9.9% to

13.8%, ash content was 20.6% to 33.5% and moisture content was 2.1% to 3.35%.

Averaged analysis data for each pair of duplicate samples (Day 1, Day 2, Day 3) were used to

prepare 50% fat materials for use in this study.

Preliminary results indicated that the average concentrations of the culture slurries

of SC, SE, SN, and SD (n=6 for each culture) ± standard error were 12.60±0.15,

12.12±0.01, 12.28±0.03, and 12.16±0.15 log10 cfu/g, respectively. Average bacterial

counts ± standard error on XLD from inoculated beef rendering samples were

10.60±0.269, 10.67±0.08, 10.76±0.04, and 10.65±0.08 log10 cfu/g, respectively (Table

3.1).

All Salmonella counts were conducted in a two-step process. Enumeration on

XLD had a lower detection limit of 1.4 log10 cfu/g. With the exception of SD, under all

treatment conditions, SC, SE, and SN were reduced to below the lower detection limit

across all thermal treatment times in inoculated beef samples. SD was detected until 60 s

(Fig. 3.2). To check for experimental error, day 1, day 2 and 3 rendering samples were re-

tested to add additional data points. Data shown in Fig. 3.2 represent n=42 for SD. The

presence of Salmonella noted at 0 and 30 s represented only 1 out of 24 samples and 1

out of 42 samples, respectively.

46

In uninoculated beef samples, SC, SE, and SN were reduced to or below the lower

detection limit across all thermal treatments. However, SD was detected at 60 s in the

uninoculated beef samples (Fig. 3.3). To check for experimental error, day1, day 2 and 3

rendering samples were re-tested to add additional data points. Data shown in Fig. 3.3

represents n=24, except at 0, 90, 240, 300, 360 and 420 s for SD. Two out of the 24

samples were determined to be positive in uninoculated beef at 60 s. A similar result was

noted at 0 s with 1 positive out of 42 samples. The uninoculated SD control sample had

Salmonella present for longer thermal treatment than the inoculated sample indicating the

presence of a background culture of either thermally resistant Salmonella or a thermally

resistant microorganism(s) that is detected as Salmonella using current methodology.

Enumeration on TSA had an upper detection limit of 4.3 log10 cfu/g. Under all

treatment conditions, bacterial plate counts on TSA for SC, SE, SN, and SD inoculated

beef samples were above the upper detection limit after all thermal treatments (Fig. 3.4).

In uninoculated beef samples used as controls for the SC, SE, SN, and SD experiments,

plate counts on TSA were above the upper detection limit after all thermal treatments in

uninoculated beef samples (Fig. 3.5).

Pre-enrichment results on RV and TT were confirmed using both latex

agglutination and ChromAgar™; the following results are reported as confirmed findings.

The unheated, inoculated controls plated on XLD after pre-enrichment in RV and TT

were significantly higher (P<0.05) than the heated, inoculated samples (this control is

indicated as unheated on Fig. 3.6, 3.7, 3.8 and 3.9). In general, Salmonella serotypes in

heated, inoculated samples declined with longer thermal treatment (Fig. 3.6 and 3.8). The

47

number of positive samples for Salmonella for each inoculated and uninoculated samples

in either RV or TT validated by the two confirmation tests are shown in Tables 3.4 and

3.5. Some samples that were reported as present had high standard errors.

In the SC samples, Salmonella was reduced to 0 at all time intervals after 0 s in

RV and TT with the exception of reappearing at 120 s in TT (Fig. 3.8). Populations of

Salmonella in the SE inoculated samples were reduced but not completely eliminated at 0

s in both RV and TT pre-enrichments (Fig. 3.6 and 3.8). For SE samples pre-enriched in

RV, Salmonella levels were reduced to 0 at 30, 60, 180, 360, and 420 s but were noted at

all other times (Fig. 3.6). Salmonella was present in SE inoculated samples at every time

interval until eliminated at 180 s and afterwards in TT pre-enrichments (Fig. 3.8).

Although populations were reduced, Salmonella was not eliminated until 360 s on SN

inoculated samples in RV pre-enrichments and until 300 s on TT pre-enrichments (Fig

3.6 and 3.8). In the SN and SD experiments, a population of Salmonella appeared to be

present in both inoculated and uninoculated samples and appeared to be more thermally

resistant than Salmonella detected on the SC and SE experiments (Fig. 3.6, 3.7, 3.8, and

3.9). In the heated, inoculated samples, SD was reduced to 0 at 30 s and 60 s, was present

at 90 s, was killed at 120 s and 180 s, and was present at 240 s in both RV and TT pre-

enrichments (Fig. 3.6 and 3.8). At 360 and 420 s, SD was reduced to 0 in RV pre-

enrichments (Fig. 3.6). At 360 s SD was present but at 420 s was reduced to 0 in TT pre-

enrichments (Fig. 3.8). Since 420 s was the maximum time tested, future studies should

include longer treatment times (Fig. 3.6 and 3.8).

Variations were noted in Salmonella populations in heated uninoculated samples

48

(Fig. 3.7 and 3.9). Salmonella was not detected at any thermal treatment time in the SC

experiments using RV enrichment but was detected at 90 s only in TT pre-enrichments

(Fig. 3.7 and 3.9). Salmonella was not detected in the SE experiments at 0, 15, 30, 60,

240, 300, 360, and 420 s in RV pre-enrichments but was detected at 90, 120, and 180 s

(Fig. 3.7). In TT, Salmonella was present in the SE experiments in all thermal treatment

times up to 180 s and was absent at 240 and 300 s (Fig. 3.9). In the SN experiments,

Salmonella was reported as in heated, uninoculated samples until 300 s in RV (Fig 3.7).

Also in the SN experiments, Salmonella was present in heated uninoculated samples until

240 s in TT (Fig. 3.9). In the SD experiments in RV, Salmonella was not detected at 0,

15, 120, 360, and 420 s in the heated, uninoculated samples (Fig. 3.7). In TT during the

SD study, Salmonella was not detected at 15, 30, 120, and 420 s (Fig. 3.9).

The estimated D values for Salmonella in beef rendering samples containing 50%

fat at 115.6ºC pre-enriched in RV and validated by two confirmation tests were

calculated. SC and SE had D values of 0.01 and 0.29 min, respectively, while SN and SD

had longer D values of 0.58 and 0.60 min (Table 3.2). The estimated D values for

Salmonella serotypes in beef rendering samples containing 50% fat at 115.6ºC pre-

enriched in TT and validated by two confirmation tests also were determined. SC and SE

had D values of 0.30 and 0.29 min, respectively, while SN and SD had D values of 0.49

and 0.70 min, respectively (Table. 3.3).

In the sieve separation experiment, each particle size fraction was indicated as a

percentage of the total weight of the rendering sample. The largest fraction of particles

collected was collected on the 25 µm mesh sieve and represented 56.6 ± 1.5% of the

49

original sample. Sieves 35, 45, 60, 80, 120, 170 and 230 µm collected 3.7± 0.3%, 4.5±

0.1%, 4.2 ± 0.3%, 3.7± 0.2%, 3.9 ± 0.8%, 4.70 ± 0.81% and 5.4 ± 0.5%, respectively

(Fig. 3.10).

Discussion

Due to the large number of samples plated per day, a preliminary experiment was

conducted to determine the percent recoveries ± standard error for each Salmonella

culture from beef rendering. The purpose of the preliminary study was to reduce plating

of each inoculated, unheated sample through extended dilutions during the study.

However, more accurate data would be obtained if plating of each inoculated, unheated

sample had been conducted. In future experiments, this control should be included.

Enumeration on XLD indicated that SC, SE and SN were reduced to below the

detection limit after the initial thermal treatment in inoculated rendering samples.

Similarly, in the uninoculated samples, SC, SE, and SN were reduced to below the

detection limit after the initial thermal treatment. SD, however, was detected at 30 s in the

inoculated samples and at 60 s in the uninoculated samples indicating the presence of a

thermally resistant bacterial strain in the background of the samples. Salmonella was

detected as present in both inoculated and uninoculated SD samples after thermal

treatment (Fig. 3.2 and 3.3). It should be noted that a positive Salmonella result from

current methodology on either inoculated or uninoculated was not validated by genetic

analysis or serotyping which would be necessary for confirmation in this study. Other

explanations for differences in recovery of Salmonella could be due to variation in

particle size distribution in the sample. SD or background organisms appearing to be

50

Salmonella in the samples may have been entrapped in a bone particle or in fat. A particle

size distribution test was conducted and showed great variability among sizes of bone

fragments. Due to the nature of rendering material collection, Salmonella could be

present in the porous structure of bone. Additionally, Salmonella could have been coated

in fat or tissue allowing for a protective effect due to slower thermal conductivity of

particles, fat and tissue. The samples in this study were randomly placed in the heating

block and, therefore, sampling error was not considered a cause for the observed

variability.

Enumeration on TSA for both inoculated and uninoculated samples indicated the

presence of the bacteria in the background of the rendering samples. The mean bacterial

counts of all samples, under all thermal treatments, were above the detection limit of 4.3

log10 cfu/g. Glenn (2006) conducted a study on the bacterial loads in raw rendering

materials, but the current study was focused on the bacterial loads in finished rendered

materials. A wide variety of heat resistant or post-process contaminating bacteria could

be present in the rendering materials; therefore, the presence of 4.3 log10 cfu/g in the

rendering samples is not unexpected.

From the preliminary study, it was determined approximately 10 log10 cfu/g of

each Salmonella culture could be recovered from inoculated rendering samples. This

concentration exceeds the detection limit of the direct plating method utilized to

enumerate on TSA. The presence of bacteria after 420 s of thermal treatment at 115.6ºC

on TSA indicated the presence of heat resistant bacteria in the background of the

rendering samples. Autoclaving requires exposure to 121ºC at 15 psi of pressure for a

51

minimum of 15 min to kill most bacteria (Laroussi and Leipold 2004). Bacterial

endospores are very heat resistant and there have been cases where endospores have not

been killed under autoclave conditions (Tuominen et al. 1994). Therefore, the thermally-

resistant bacteria in the background of rendering materials could potentially be spore-

forming bacteria. The design of this experiment did not allow for further analysis of these

heat-resistant bacteria. However, future experiments will isolate and identify these

bacterial species through genetic analysis or serotyping.

Results of RV and TT pre-enrichments indicated variation in recovery amongst

cultures identifying as Salmonella in the SC, SE, SN, and SD inoculated and

uninoculated samples. In inoculated and uninoculated samples pre-enriched in RV, the

presence of SC or organisms appearing to be Salmonella declined after the application of

heat. In TT, the presence of SC or microbes appearing to be Salmonella followed a

similar trend as the RV pre-enriched samples. However, Salmonella were detected in

both inoculated and uninoculated samples at 90 and 120 s in TT. In RV, SE or bacteria

detected as Salmonella were present in both inoculated and uninoculated samples at 90

and 120 s. However, in TT, SE or organisms presenting as Salmonella were detected in

both inoculated and uninoculated samples at 0, 15, 30, 60, 90, and 120 s. The presence of

Salmonella or organisms detected as Salmonella at 90 s and 120 s may be background

bacteria. The presence of SN or organisms detected as Salmonella were present at 0, 15,

30, 60, 90, 120, 180, 240 s in both RV pre-enriched inoculated and uninoculated samples.

SN or Salmonella-like bacterial species were detected in TT until 300 s in inoculated and

uninoculated samples. Positive results in inoculated samples may be due to background

52

organisms. SD or organisms detected as Salmonella were present in both inoculated and

uninoculated samples pre-enriched in RV at 60, 90, 240 and 300 s. In TT, SD or

organism detected as Salmonella were present at 0, 90, 240, 300 and 360 s in inoculated

and uninoculated samples. Again, positive results in inoculated samples may be due to

background organisms. Another explanation for the results of this study could be that

Salmonella species may have been entrapped in bone particles or in fat. In comparing the

presence of Salmonella in inoculated samples pre-enriched in either RV or TT, the

presence of Salmonella or a Salmonella-like organism appeared to follow similar trends

across all experiments.

The presence of a thermally resistant organism reacting as Salmonella has been

well-noted in the rendering samples in this study. The rendering process recycles inedible

animal tissue to produce products that can be used in animal feed. Therefore, it is

hypothesized that an unknown bacterial strain(s) may have acquired thermal resistance

and/or Salmonella-like characteristics through repetitive cycles of animal feed, animals

and rendering. Inedible animal tissues including the gastrointestinal tract and its inherent

microorganisms would be rendered and the cycle through animal feed to animal to

slaughter to rendering could hypothetically repeat. Potentially these conditions could

select for thermally resistant microorganisms. Since this hypothesis has not been tested, it

is vital that this unknown strain or strains is isolated in future experimentation to

determine its identity and characteristics.

Preliminary estimated D values were calculated. SN and SD appeared to have

longer D values than SC and SE. As a general rule of thumb, with increase in

53

temperature, the thermal lethality increases (Earle and Earle 1983). Liu et al. (1969)

reported D values for Salmonella senftenberg 775 W were highly variable between 10 to

115 min at 70ºC in meat and bone meal. Lui et al. (1969) conducted their study in meal

and the current study was conducted in cooked beef rendered products containing 50%

fat content. Similar to the Lui et al. (1969) study, the D values of this study were variable

and high which could potentially be due to the thermally resistant background

organism(s).

Further research needs to be conducted at 240ºF (115.6ºC) for longer time

intervals to ensure that SC, SE, SN and SD are destroyed. It should be noted the results of

this study were obtained from the lower end of the cooking temperatures utilized in the

rendering industry. Many rendering facilities process materials at higher temperatures

close to 280ºF (137.8ºC) to 290ºF (143.3ºC) for 40 to 90 min in order to produce

microbiologically safe products (Meeker and Hamilton 2006). However, the industry also

employs a different type of cooker known as a Carver-Greenfield unit. These units

operate at lower temperatures, typically closer to 240ºF (115.6ºC). Carver-Greenfield

units operate under vacuum to process the materials at this lower temperature (Meeker

and Hamilton 2006).

It was necessary to grind rendering materials for transfer into stainless steel tubes.

Factors for comparing data to typical bone particle sizes will necessary for future

experiments. Thermal conductivity studies on larger bone particles could provide further

understanding of thermal lethality in rendering materials.

54

Figure 3.1. Method utilized to report RV and TT pre-enrichments results on XLD

validated by two confirmation tests at each thermal treatment. If both plates were

negative, the result was assigned a 0 (A). If one was positive and one was negative, the

result was assigned a 0.5 (B). If both were positive, the result was assigned a 1.0 (C).

A. B. C.

- - + - + +

0.0 0.5 1.0

55

Figure 3.2. Enumeration of Salmonella on XLD from beef rendering samples (50% fat)

inoculated with Salmonella Choleraesuis (SC), Salmonella Enteriditis (SE), Salmonella

Newport (SN), and Salmonella Dublin (SD).1

1The lower limit of detection is 1.4 log10 cfu/g of Salmonella (n=24, except for SN and

SD n=42 at 0, 90, 240, 300, 360 and 420 s).

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Unheated

0 15

30

60

90

120

180

240

300

360

420

Mea

n B

acte

rial

Count,

log

10 c

fu/g

Thermal Time Treatment, s

SC

SE

SN

SD

Inoculated

56

Figure 3.3. Enumeration of Salmonella on XLD from uninoculated beef rendering

samples (50% fat) for Salmonella Choleraesuis (SC), Salmonella Enteriditis (SE),

Salmonella Newport (SN), and Salmonella Dublin (SD).1


SD n=42 at 0, 90, 240, 300, 360 and 420 s).

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Unheated

0 15

30

60

90

120

180

240

300

360

420

Mea

n B

acte

rial

Count,

log

10 c

fu/g

Thermal Treatment Time, s

SC

SE

SN

SD

Uninoculated

57

Figure 3.4. Enumeration of total bacteria on TSA from beef rendering samples (50% fat)

inoculated with Salmonella Choleraesuis (SC), Salmonella Enteriditis (SE), Salmonella

Newport (SN), and Salmonella Dublin (SD).1


SD n=42 at 0, 90, 240, 300, 360 and 420 s).

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Unheated

0 15

30

60

90

120

180

240

300

360

420

Mea

n B

acte

rial

Count,

log

10 c

fu/g


SC

SE

SN

SD

Inoculated

58

Figure 3.5. Enumeration of total bacteria on TSA from uninoculated beef rendering




SD n=42 at 0, 90, 240, 300, 360 and 420 s).

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Unheated

0 15

30

60

90

120

180

240

300

360

420

Mea

n B

acte

rial

Count,

log

10 c

fu/g


SC

SE

SN

SD

Uninoculated

59

Figure 3.6. Presence or absence ± standard deviation of Salmonella for each Salmonella

Choleraesuis (SC), Salmonella Enteriditis (SE), Salmonella Newport (SN), and

Salmonella Dublin (SD) inoculated, RV pre-enriched beef rendering samples (50% fat).1

1A count of 0 represent the absence of Salmonella, while a count of 1 represents the

presence of Salmonella (n=24, except for SN and SD n=42 at 0, 90, 240, 300, 360 and

420 s).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Unheated

0 15

30

60

90

120

180

240

300

360

420

Pre

sence

/ A

bse

nce

Thermal Treatment Time (s)

SC

SE

SN

SD

Inoculated

60

Figure 3.7. Presence or absence ± standard deviation of Salmonella for each RV pre-

enriched, uninoculated beef rendering samples (50% fat) for Salmonella Choleraesuis

(SC), Salmonella Enteriditis (SE), Salmonella Newport (SN), and Salmonella Dublin

(SD).1



420 s)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Unheated

0 15

30

60

90

120

180

240

300

360

420

Pre

sence

/ A

bse

nce


SC

SE

SN

SD

Uninoculated

61



Salmonella Dublin (SD) inoculated, TT pre-enriched beef rendering samples (50% fat).1



420 s).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Unheated

0 15

30

60

90

120

180

240

300

360

420

Pre

sence

/ A

bse

nce


SC

SE

SN

SD

Inoculated

62

Figure 3.9. Presence or absence ± standard deviation of Salmonella for each TT pre-

enriched, uninoculated beef rendering samples (50% fat) for Salmonella Choleraesuis


(SD).1



420 s).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Unheated

0 15

30

60

90

120

180

240

300

360

420

Pre

sence

/ A

bse

nce


SC

SE

SN

SD

Uninoculated

63

Figure 3.10. Mean percent particle size distribution ± standard error of beef rendering

samples collected from a rendering plant on three different days (n=30). Each fraction of

particle size was indicated as a percentage of the total weight of the rendering sample.

The error bars indicate standard error for each data point.

0

10

20

30

40

50

60

70

25 35 45 60 80 120 170 230

Mea

n P

erce

nta

ge

of

Ori

gin

al S

ample

(%

)

Standard Sieve Sizes (μm)

64

Table 3.1. Salmonella Choleraesuis (SC), Salmonella Enteriditis (SE), Salmonella

Newport (SN) and Salmonella Dublin (SD) after inoculation into beef rendering materials

and plated onto XLD (n=6).

Serotype

Average Broth Culture,


Average in Beef Samples,


SC 12.60±0.15 10.60±0.29

SE 12.12±0.01 10.67±0.08

SN 12.28±0.03 10.76±0.04

SD 12.16±0.15 10.65±0.08

65

Table 3.2. Estimated D values for Salmonella Choleraesuis (SC), Salmonella Enteriditis

(SE), Salmonella Newport (SN), and Salmonella Dublin (SD) in beef rendering samples

(50% fat) at 115.6ºC pre-enriched in RV and validated by two confirmation tests.

Serotype Estimated D Value, min

SC 0.01

SE 0.29

SN 0.58

SD 0.60

66


(SE), Salmonella Newport (SN), and Salmonella Dublin (SD) in beef rendering samples

(50% fat) at 115.6ºC pre-enriched in TT and validated by two confirmation tests.


SC 0.30

SE 0.29

SN 0.49

SD 0.70

67

Table 3.4. Number of samples positive for Salmonella in Salmonella Choleraesuis (SC),

Salmonella Enteriditis (SE), Salmonella Newport (SN) and Salmonella Dublin (SD)

inoculated beef rendering samples (50% fat) after pre-enrichment in RV or TT broth

(n=24, except for SN and SD n=42 at 0, 90, 240, 300, 360 and 420 s).

RV TT

Serotypes Thermal

Treatment

Time, s

Number of Positive Samples Number of Positive Samples

SC Unheated 24 out of 24 samples 24 out of 24 samples

0 0 out of 24 samples 0 out of 24 samples









SE Unheated 24 out of 24 samples 24 out of 24 samples










SN Unheated 42 out of 42 samples 42 out of 42 samples












SD Unheated 42 out of 42 samples 42 out of 42 samples


68











69

Table 3.5. Number of samples positive for Salmonella in uninoculated beef rendering


Salmonella Newport (SN) and Salmonella Dublin (SD) after pre-enrichment in RV or TT

broth (n=24, except for SN and SD n=42 at 0, 90, 240, 300, 360 and 420 s).

RV TT

Serotype Thermal

Treatment

Time, s

Number of Positive Samples Number of Positive Samples

































70













71

References


R.M. 2011. Chapter 5 Salmonella [online]. In: FDA bacteriological analytical

manual (BAM). AOAC International, Gaithersburg, M.D.


htm (Accessed 9 April 2013).


and its relationship to human foodborne illness. Clin. Infect. Dis. 35:859-865.

Earle, R.L., and M.D. Earle. 1983. Unit operations in food processing, web edition. The

New Zealand Institute of Food Science and Technology, Inc. New Zealand.

Feng, P. 2001. Rapid methods for detecting foodborne pathogens. In: FDA

bacteriological analytical manual (BAM). AOAC International, Gaithersburg,

MD.


htm (accessed 9 April 2013).


poultry rendering materials. MS thesis. Clemson Univ. Clemson, SC.

Heldman, D., and R. W. Hartel (ed.). 1998. Principles of food processing. Aspen

Publishers, Gaithersburg, MD.

Jay, J.M. 2005. Modern food microbiology. 6th ed. Aspen Publishers Inc., Gaithersburg,

MD.

Laroussi, M., and F. Leipold. 2004. Evaluation of the roles of reactive species, heat, and

UV radiation in the inactivation of bacterial cells by air plasmas at atmospheric

pressure. Int. J. Mass Spectrom. 223: 81-86.

Liu, T.S., G.H. Snoeyenbos, and V.L, Carlson. 1969. Thermal resistance of Salmonella

senftenberg 775W in dry animal feeds. Avian Dis. 13: 611-631.

Meeker, D.L., and C.R. Hamilton. 2006. An overview of the rendering industry. In: D.L.

Meeker (ed.), Essential rendering all about the animal by-products industry.

National Renderers Association, Arlington, VA. p. 1-16.

Teixeira. A.A. 2006. Simulating thermal food processes using deterministic models. In:

D. Sun (ed.), Thermal food processing. CRC Press, Taylor & Francis Group, Boca

Raton, FL. p. 73-106.

72

Tuominen, L., T. Kairesalo, and H. Hartikainen. 1994. Comparison of methods for

inhibiting bacterial activity in sediment. Appl. Environ. Microbiol. 60: 3454–

3457.

United States Food and Drug Administration (FDA). 2010. Compliance policy guide Sec.

690.800 Salmonella in animal feed. Draft guidance. United States Food and Drug

Administration. Rockville, MD. http://www.gpo.gov/fdsys/pkg/FR-2010-08-

02/pdf/2010-18873.pdf (Accessed 9 May 2013).




policy-guide-sec-690800-salmonella (Accessed 9 August 2013).

73

CHAPTER 4

VALIDATION OF THERMAL DESTRUCTION OF SALMONELLA IN RENDERED

POULTRY PRODUCTS

Abstract

Only a portion of a food animal is considered edible by humans. The remainder of

the animal tissue is considered inedible and typically rendered into animal co-products.

Rendering recycles the residual animal tissue from food animals into stable, value-added

materials for use primarily in animal feeds. Therefore, the rendering industry must have

validation data on the thermal lethality of rendering thermal process to ensure the

destruction of animal disease pathogens in finished products. The unique high fat, bone

and protein content of rendering materials leaves the industry with no comparable

thermal death time values from the human food industry or any other industry. The

objective of this study is to determine thermal death time values for poultry rendering

materials containing 50% fat content for four pathogenic Salmonella recognized by FDA

as hazardous for animal feeds (Salmonella Choleraesuis (SC), Salmonella Enteritidis

(SE), Salmonella Newport (SN) and Salmonella Dublin (SD)). Recoverability of

Salmonella varied after pre-enrichment in either RV or TT broth. Levels of Salmonella in

the samples did not exhibit a straight line decrease with increasing thermal treatment

times. In thermal treatment trials extended up to 420 s at 240ºF (115.6ºC), Salmonella

were detected in the SC, SE, SN and SD samples at 360 s. Thermally resistant Salmonella

or Salmonella-like strains in the background were detected up to 360 s of treatment in

uninoculated controls. Future experiments will be needed to validate whether these

organisms are Salmonella.

74

Introduction

Rendered animal products can potentially be contaminated with Salmonella spp.

Approximately 85% of rendered products are used as animal feed ingredients which can

potentially transmit Salmonella to humans through the food chain (Crump et al. 2002).

Loken et al. (1968) tested 1,395 rendered products from seven different plants and

detected the presence of Salmonella in 241 (17%) of the samples. The study also tested

the plant via environmental swabs, and Salmonella was isolated from 359 out of 1901

(19%) of the swabs. In a study conducted in 1977, Salmonella was detected in 81% of the

meat meal and 40% of the feather meal produced over a four mo period in Ontario feed

mills (Hacking et al. 1977). In 1993 and 1994, FDA conducted two separate studies

examining rendered animal feed products for the presence of Salmonella enterica and

determined 56% and 25% of the samples, respectively, were positive (McChesney et al.,

1995; Crump et al., 2002). Troutt et al. (2001) examined 17 rendering facilities located in

seven midwestern states of the United States. No Salmonella was found in crax samples

or in the rendering processing environment. However, the finished rendered products

contained 12 serovars of Salmonella. Franco (2005) reported Salmonella cells were

present in low numbers in animal feed after analyzing approximately 200 rendered

animal protein meal samples over a 12 mo period. Kinley et al. (2009) examined products

from 12 rendering facilities in the United States and detected 13 Salmonella serovars. In

2010, Kinley et al. determined the prevalence of Salmonella and Enterococcus spp. in

poultry meal or feather meal from 12 United States rendering companies. Enterococcus

75

spp. were detected in 81.3% of the samples and accounted for up to 54% of the total

bacterial counts in some samples. Salmonella was only detected in 8.7% of the samples.

To ensure the microbiological safety of rendering products, rendering facilities

utilize thermal processing for 40 to 90 min at 240 to 290ºF (115.6 to 143.3ºC) (Meeker

and Hamilton, 2006). Marginal processing conditions potentially could result in microbial

survival (Crump et al., 2002). Thermal death time (TDT) is a factor of time, temperature,

material matrix and organism (Heldman and Hartel, 1998). Decimal reduction time (D

value) indicates the time required for a one log10 reduction of a particular organism at a

specific temperature (Heldman and Hartel, 1998). TDT of Salmonella has been

investigated in food products (Murphy et al., 2000; D’Aoust, 2001; Murphy et al., 2004;

Bucher et al., 2008), but few studies have been conducted in rendered animal products.

Franco (1997 and 2005) conducted surveys of Salmonella in rendered animal co-products

and suggested rendering processes destroy Salmonella. Ramirez-Lopez (2006) studied

TDT of a single unknown isolate from animal co-products. However, data has never been

generated on TDT of Salmonella in rendered poultry materials. Since this factor must

consider the parameters of matrix, temperature and organism, it was necessary to conduct

validation in the actual rendering material matrices. The objective of this study was to

determine the TDT and D values for four pathogenic Salmonella recognized by FDA as

hazardous for animal feeds (Salmonella Choleraesuis, Salmonella Enteritidis, Salmonella

Newport and Salmonella Dublin) in poultry rendering materials containing 50% fat

content (FDA, 2010; FDA, 2013) at 240ºF (115.6ºC).

76



Samples of poultry crax and poultry fat were obtained from a southeastern

rendering company on three separate days. Crax is a solid material composed of protein,

minerals, and residual fat that is discharged from the screw press during the rendering

process and is typically further ground into meat and bone meal (Meeker and Hamilton,

2006). The crax samples were submitted in duplicate to Clemson University Agricultural

Service Laboratory for ash, fat, and moisture content analysis. The fat and crax samples

were mixed to produce 50% fat samples. A food processor bowl, blade and lid were

disinfected by rinsing in Antibac B™ (Diversey Corporation, Cincinnatti, OH) dissolved

in distilled deionized water (ddH2O) (0.6 g per L) for approximately 2 min, followed by

rinsing 5 times with sterile ddH2O. Particle size was reduced by processing for

approximately 10 min on the pulse setting in the disinfected food processor (Robot

Coupe Model R2 Ultra, Ridgeland, MS) prior to conducting the experiments. A sterile

stainless steel spatula was used to scrape material from the sides during pauses in

processing. All samples were stored under refrigeration until needed for experimentation.


Four pathogenic Salmonella serotypes recognized by FDA as hazardous for

animal feeds (Salmonella Choleraesuis (FDA 8326) (SC), Salmonella Enteritidis (USDA

H4386) (SE), Salmonella Newport (USDA H1073) (SN) and Salmonella Dublin (FDA

23742) (SD)) were obtained for this study (FDA, 2010; FDA, 2013). SE and SN were

obtained from Dr. Vijay Jejuna of the USDA Agricultural Research Service, Microbial

77

Food Safety Research Unit, 600 East Mermaid Lane, Room 2129, Wyndmoor, PA 19038.

SC and SD were obtained from the food microbiology culture collection from

collaborator Dr. Xiuping Jiang at Clemson University.






extract was chosen as the best media based highest cell densities determined from optical

density measurements (µQuant Universal Microplate Spectrophotometer, Bio-Tek

Instruments, Winooski, VT) at 600 nm and dilution plating in duplicate onto onto

bismuth sulfite agar (90003-904, VWR Scientific Products, Suwanee, GA), Hektoen

enteric agar (9004-054, VWR Scientific Products), xylose lysine deoxycholate (XLD)

(90003-996, VWR Scientific Products), and trypticase soy agar (TSA) (90000-050, VWR

Scientific Products).

As a preliminary study, each individual Salmonella serotype was plated onto

bismuth sulfite agar, Hektoen enteric agar, XLD, and TSA. Enumeration data indicated

use of XLD and TSA as the preferred agar media for enumerating SC, SE, SN, and SD.

A preliminary goal of this experiment was to obtain concentrated bacterial slurry

to use in inoculating poultry rendering materials for thermal processing. The average

concentrations of cells in broth for SC, SE, SN, and SD after 24 h incubation at 35ºC

were 8.66±0.02, 8.56±0.03, 8.80±0.06, and 8.65±0.03 log10 cfu/g, respectively.

78

Preliminary experiments were conducted to determine the volume of culture as well as

concentration rate necessary. Enumeration on XLD and TSA verified that 5 L of a 24 h

Salmonella culture grown in TSB with 0.1% (wt/vol) yeast extract and then concentrated

by centrifugation was optimal. Centrifugation was conducted at 7,000 x g for 7 min (GSA

rotor, DuPont RC5C Sorvall Instruments Centrifuge, DuPont Company, Newtown, CT)

at 4ºC in sterile centrifuge bottles (47735-696, VWR Scientific Products, Suwanee, GA)

and the supernatant was discarded after autoclaving. The pellet was resuspended in 5 mL

sterile TSB. In preliminary studies conducted 3 times in duplicate (n=6), the average

bacterial slurry concentrations for SC, SE, SN, and SD were 12.60±0.15, 12.12±0.01,

12.28±0.03, and 12.16±0.15 log10 cfu/g, respectively. This procedure was used to prepare

the bacterial cultures used throughout the experiment.

Each Salmonella slurry, prepared as above, was inoculated into poultry rendering

material at the rate of 100 µL culture per 1 g sample. A preliminary study was conducted

to determine the difference in mean bacterial counts of the inoculated samples versus the

bacterial slurry in TSB with 0.1% (wt/vol) yeast extract using two different methods.

Method 1 was the serial dilution of each broth culture as well as each inoculated sample

to 10-14

utilizing the standard Class O phosphate/magnesium chloride dilution buffer

(Wehr and Frank, 2004). Method 2 was the serial dilution of each broth culture and

sample to 10-14

using pre-warmed (32ºC) modified Class O phosphate/magnesium

chloride diluent. Controls included media and uninoculated poultry rendering samples.

Each experiment was conducted 3 times in duplicate (n=6).

79

Thermal Death Time Trials

Stainless steel sample tubes (8.5 cm length, 1.6 cm outer diameter, 1.3 cm inner

diameter) were custom manufactured by a local company by boring 304 stainless steel

rods. These tubes were capped (60825-801, VWR International, Suwanee, GA) and

autoclaved. Poultry rendering samples (50% fat) were aseptically transferred (1 g) into

sixteen sterile tubes. The tubes were placed in an analog dry block heater (Model#12621-

108, VWR International, Suwanee, GA) equipped with Model#13259-162 heating blocks

(VWR International, Suwanee, GA) set to 115.6ºC. Four of the tubes were randomly

selected as temperature controls using dial thermometers (61159-409, VWR Scientific

Products, Suwanee, GA). The tubes were heated to an internal treatment temperature of

115.6ºC prior to addition of the cultures. Each individual culture (100 µL) was directly

pipetted into 1 g of the heated rendering samples. After culture inoculation, the sample

was pipetted up and down approximately 4 times to thoroughly mix. Upon inoculation

and mixing, time measurements (0, 15, 30, 60, 90, 120, 180, 240, 300, 360, and 420 s)

began on the thermal treatment. Samples were placed on ice immediately after thermal

treatment. Additional sample tubes containing poultry rendering used for unheated

controls were placed on ice until utilized for plating. All samples were processed for

microbial content immediately after conclusion of heat treatments.

A preliminary experiment was conducted to validate the use of 1 g of sample pre-

enriched in 5 mL of sterile universal pre-enrichment broth (UPB) (95021-036, VWR

Scientific Products, Suwanee, GA) in comparison to 1 g of sample pre-enriched in 9 mL I

of UPB as per recommendations in the FDA Bacteriological Analytical Manual (BAM)

80

(Andrews et al., 2011). The stainless steel tubes used in this experiment would not hold 1

g of sample pre-enriched in 9 mL of UPB. Results indicated that the 1:5 ratio of sample

to pre-enrichment broth was as effective as the 1:9 ratio of sample to pre-enrichment

broth. Therefore, this procedure was used throughout the experiment.

Once 5 mL of sterile UPB was aseptically pipetted into each tube, the wooden

shaft of a sterile cotton-tipped applicator (89133-814, VWR Scientific Products,

Suwanee, GA) was used to thoroughly mix the sample for 30 s. Each UPB diluted sample

(0.1 mL) was directly pipetted onto XLD and TSA plates and spread using an alcohol-

flamed bent glass rod. As a control, each Salmonella slurry was serially diluted to 10-12

in

the standard Class O phosphate/magnesium chloride dilution buffer and either 1.0 mL or

0.1 mL was spread plated onto XLD and TSA. Media and dilution buffer controls also

were conducted. All plates were incubated overnight at 35ºC. In this experimental design,

XLD selected for Salmonella spp. while TSA measured total bacterial counts (aerobic,

mesophilic), which included any background bacteria and in test samples background

bacteria plus inoculated Salmonella. For each inoculate or uninoculated poultry rendering

sample, dilutions were carried out such that the direct plating on XLD and TSA had a

lower detection limit of 1.4 log10 cfu/g.

Because the direct plate counting method had a lower detection limit of 1.4 log10

cfu/g, an additional experiment was conducted in accordance with FDA Bacteriological

Analytical Manual (BAM) procedures; this second experiment had a detection limit of 1

cfu/g (Andrews et al., 2011). The remaining UPB diluted sample in the stainless steel

tube was incubated overnight at 35ºC and then vortexed (Super Mixer, 1290, Labline

81

Instruments, Inc., Melrose Park, IL) on the fast setting for approximately 30 s. The

sample (0.1 mL) was aseptically pipetted to Rappaport-Vassiliadis (RV) pre-enrichment

broth (10 mL) (95039-382, VWR Scientific Products, Suwanee, GA). The same sample

(1 mL) was aseptically pipetted to tetrathionate broth (TT) (10 mL) (90000-008, VWR

Scientific Products, Suwanee, GA). Controls included bacterial slurry and sterile media.

The samples and controls were incubated overnight at 42ºC. A 3 mm inoculation loop of

each pre-enriched sample and control was streaked onto XLD. All plates were incubated

overnight at 35ºC. Results indicated the presence or absence of Salmonella in the

samples.

As per FDA BAM recommendations to validate positive samples obtained from

the RV and TT pre-enrichments, two confirmation tests were conducted (Feng, 2001).

Latex agglutination tests (FT0203, Thermo Fisher Scientific, Waltham, MA 02454) and

ChromAgar™ (90006-158, VWR Scientific Products, Suwanee, GA) were conducted

using the each Salmonella culture as a control (BD Diagnostics, 2008; Oxoid Limited,

2013). In order to analyze the data, when duplicate results from the pre-enriched samples

were both negative the data was reported as 0.0 (Figure 4.1). If one duplicate was positive

and one was negative, it was reported as 0.5. If both duplicates were positive, it was

reported as 1.0 (Figure 4.1).

Determination of Estimated D Values

The direct plate count of each concentrated Salmonella slurry and the time at

which each culture was destroyed were compared on a graph. In a preliminary

experiment, percent recoveries of Salmonella from inoculated poultry samples were

82

calculated for each recoverable Salmonella population density. Due to the experimental

design, the actual population count from poultry rendering material was not conducted.

However, the total count in each bacterial slurry was measured. This population count

was used in estimated D value calculations. The final time the population was no longer

detected was used as the thermal death time. These data were graphed and the slope of

the line was used to calculate the estimated the D value.

Results

Analysis of poultry rendering materials indicated mean fat content was

15.97±1.13%, mean ash content was 10.55±1.14% and mean moisture content was

3.73±0.33%. Averaged analysis data for each pair of duplicate samples (Day 1, Day 2, Day

3) were used to prepare 50% fat materials for use in this study.

Preliminary results indicated that the average concentration of the culture slurries

of SC, SE, SN, and SD were 12.60±0.15, 12.12±0.01, 12.28±0.03 and 12.16±0.15 log10

cfu/g, respectively. The mean bacteria counts ± standard error on XLD from inoculated

poultry rendering materials were 10.47±0.20 15, 10.59±0.23, 10.43±0.22 and 10.40±0.13

log10 cfu/g, respectively (Table 4.1).

All Salmonella counts were conducted in a two-step process. Enumeration on

XLD had a lower detection limit of 1.4 log10 cfu/g. Under all treatment conditions, SC,

SE, SN and SD were reduced to or below the lower detection limit after initial thermal

treatment (0 s) in inoculated poultry samples (Figure 4.2). Salmonella levels were

reduced to or below lower detection limit during after initial thermal treatment (0 s) in

uninoculated poultry control samples (Figure 4.3).

83

Enumeration on TSA had an upper detection limit of 4.3 log10 cfu/g. Under all

treatment conditions, total bacterial counts in the SC, SE, SN, and SD trials were above

the upper detection limit after all thermal treatments in inoculated poultry samples

(Figure 4.4). Total bacterial counts were above the upper detection limit after all thermal

treatments in all uninoculated poultry samples (Figure 4.5).

Pre-enrichment results on RV and TT were confirmed using both latex

agglutination and ChromAgar™; the following results are reported as confirmed findings.

In general, Salmonella serotypes in heated inoculated samples declined with longer

thermal treatment (Figure 4.6 and 4.8). The positive counts for Salmonella in each

inoculated and uninoculated sample in either RV or TT validated by the two confirmation

tests are shown in Tables 4.4 and 4.5. Populations of Salmonella in the SC inoculated

samples were reduced, but did not appear to be eliminated until 360 s in RV pre-

enrichments. Although populations were reduced, Salmonella levels did not appear to be

destroyed until 420 s in TT pre-enriched, SC inoculated samples (Figure 4.6 and 4.8). In

RV, Salmonella in the SE inoculated samples was present at every time interval until it

appeared to be eliminated at 420 s (Figure 4.6). Populations of Salmonella in the SE

inoculated samples in TT were reduced to 0 at 90 s, were present at 120, 180, 240, 300

and 360 s, and appeared to be killed at 420 s (Figure 4.8). Levels of Salmonella in the SN

inoculated samples were reduced to 0 at 120 s, were present at 180 s, but were eliminated

at 240 s in RV (Figure 4.6). In TT, Salmonella populations were reduced to 0 at 120 s,

but were present again until 420 s in SN inoculated samples (Figure 4.8). For SD samples

pre-enriched in RV, Salmonella levels decreased until reaching 0 at 90, 120, 180 and 240

84

s, but Salmonella was present at 300 s on RV pre-enrichments. Salmonella appeared to be

eliminated at 360 s and thereafter (Figure 4.6). In TT, Salmonella was reduced to 0 at 300

s, was present at 360 s and appeared to be killed at 420 s in SD inoculated samples

(Figure 4.8). Since 420 s was the maximum time tested, future studies should include

longer treatment times (Figure 4.6 and 4.8).

Variations were noted in Salmonella populations in heated uninoculated samples

(Figure 4.7 and 4.9). Salmonella was detected at 0, 60, 90, 120, 180, 240 and 360 s in RV

pre-enrichments for SC uninoculated samples but was not detected at 15, 30, 300, and

420 s (Figure 4.7). In TT, Salmonella levels in the uninoculated controls for SC were not

reduced to 0 until 420 s (Figure 4.9). In RV and TT pre-enrichments for the uninoculated

SE samples, populations of Salmonella were present in all thermal treatment times up to

420 s (Figure 4.7 and 4.9). For the uninoculated SN samples, Salmonella was present at

15, 30, 60, 90, 180, 240, and 300 s in RV (Figure 4.7). In TT, Salmonella was not

detected in uninoculated SN controls at 0, 120, 360, and 420 s (Figure 4.9). Levels of

Salmonella in uninoculated SD samples were not detected at 60, 90, 120, 180, 360 and

420 s in RV (Figure 4.7). In TT, Salmonella was not detected at 0 s and 420 s but was

present at all other thermal treatment times in the uninoculated SD samples (Figure 4.9).

The estimated D values for Salmonella in poultry rendering samples containing

50% fat at 115.6ºC pre-enriched in RV and validated by two confirmation tests were

calculated. SC, SE, SN, and SD had D values of 0.60, 0.67, 0.39, and 0.58 min,

respectively (Table 4.2). The estimated D values for Salmonella serotypes in poultry

rendering samples containing 50% fat at 115.6ºC pre-enriched in TT and validated by

85

two confirmation tests also were determined. SC, SE, SN, and SD had D values of 0.70,

0.67, 0.67, and 0.67 min, respectively (Table 4.3).

Discussion

Since large numbers of samples were plated per day, a preliminary experiment

was conducted to determine the percent recoveries ± standard error for each Salmonella

culture from poultry rendering instead of conducting a full dilution series on each day of

plating. The preliminary study allowed for the reduction of plating of each inoculated,

unheated sample through extended dilutions during the study. However, future

experiments should be designed to conduct the plating of each inoculated, unheated

sample to obtain more accurate data.

Enumeration on XLD indicated that SC, SE, SN, and SD were reduced to below

the detection limit after the initial thermal treatment in inoculated and uninoculated

rendering samples (Figures 4.2 and 4.3). The presence of the bacteria in the background

of the rendering samples was indicated through enumeration on TSA for both inoculated

and uninoculated samples (Figures 4.4 and 4.5). The mean bacterial counts of all

samples, under all thermal treatments, were above the detection limit of 4.3 log10 cfu/g.

The current study enumerated total bacterial content in finished rendered materials.

However, Glenn (2006) conducted a study on the bacterial loads in raw rendering

materials and detected high levels of microbial content. Diverse populations of non-

pathogenic and pathogenic heat-resistant bacteria could be contaminants in rendering

materials due to either survival of the rendering cooking process or post-process

contamination. Therefore, the presence of 4.3 log10 cfu/g in the rendering samples is not

86

unexpected. After thermal treatments of 420 s at 115.6ºC, bacterial populations were still

present as measured on TSA indicating the presence of heat resistant bacteria in the

background of the rendering samples. Autoclaving requires exposure to 121ºC at 15 psi

of pressure for a minimum of 15 min to kill most bacteria (Laroussi and Leipold 2004).

Bacterial endospores are very heat resistant and in certain cases are not killed under

autoclave conditions (Tuominen et al. 1994). Therefore, the thermally-resistant bacteria

in the background of rendering materials could potentially be spore-forming bacteria. The

design of this experiment did not allow for further analysis of these heat-resistant

bacteria. However, future experiments are needed to isolate and identify these bacterial

species through genetic analysis or serotyping.

Results of RV and TT pre-enrichments indicated variation in recovery of

Salmonella amongst SC, SE, SN, and SD inoculated and uninoculated samples. SC or

organisms detected as Salmonella were present in both inoculated and uninoculated

samples pre-enriched in RV and TT but it appeared more frequently in TT pre-enriched

samples. In RV and TT, SE or bacteria detected as Salmonella were present in both

inoculated and uninoculated samples at all thermal treatment times up to 420 s, except in

inoculated samples pre-enriched in TT at 90 s. The presence of SN or organisms detected

as Salmonella peaked at 90 s, decreased to 0 at 120 s, and re-emerged at 180 s in both RV

and TT pre-enriched inoculated samples. SN or a Salmonella-like bacterial species was

detected in uninoculated samples pre-enriched in TT at 90 s, not detected at 120 s, and

detected again at 180 s. This trend was also observed in uninoculated samples pre-

enriched in RV. SD or organisms detected as Salmonella were present in both inoculated

87

and uninoculated samples pre-enriched in RV and TT but it appeared more frequently in

TT pre-enriched samples. Positive results in inoculated samples may be due to

background organisms. It should be noted that a positive Salmonella result from current

methodology on either inoculated or uninoculated was not validated by genetic analysis

or serotyping which would be necessary for confirmation in this study.

Due to the nature of rendering material collection, Salmonella could be present in

the porous structure of bone. Additionally, Salmonella could have been coated in fat or

tissue allowing for a protective effect due to slower thermal conductivity of bone

particles, fat and/or tissue. The samples in this study were randomly placed in the heating

block and therefore, this factor was not considered a cause for the observed variability.

The presence of a thermally resistant organism reacting as Salmonella has been

well-noted in the rendering samples in this study. The rendering process recycles inedible

animal tissue to produce products that can be used in animal feed. Therefore, it can be

hypothesized that an unknown bacterial strain(s) may have acquired thermal resistance

and/or Salmonella-like characteristics through repetitive cycles of rendered animal feed

to animals to rendering. In this hypothesis, inedible animal tissues including the

gastrointestinal tract and its inherent microbes would be rendered and the cycle through

animal feed to animal to slaughter to rendering would repeat. These conditions

potentially could select for thermally resistant microbes. Since this hypothesis has not

been tested, it is vital that this unknown strain(s) is isolated in future experimentation to

determine its identity and characteristics.

88

Preliminary estimated D values were calculated. In general, with increase in

temperature, thermal lethality increases (Earle and Earle 1983). Liu et al. (1969) reported

D values for Salmonella senftenberg 775 W were highly variable between 10 to 115 min

at 70ºC in meat and bone meal. Lui et al. (1969) conducted their study in meal and the

current study was conducted in cooked poultry rendered products containing 50% fat

content. Similar to the Lui et al. (1969) study, the D values of this study were variable

and high which could potentially be due to the thermally resistant background

organism(s).

Further research needs to be conducted at 115.6ºC for longer time intervals to

ensure that SC, SE, SN and SD are destroyed. It should be noted the results of this study

were obtained from the lower end of the cooking temperatures utilized in the rendering

industry. Many rendering facilities process materials at higher temperatures closes to

280ºF (137.8ºC) to 290ºF (143.3ºC) for 40 to 90 min in order to produce

microbiologically safe products (Meeker and Hamilton 2006). However, the industry also

employs a different type of cooker known as a Carver-Greenfield unit. These units

operate under vacuum at lower temperatures, typically closer to 240ºF (115.6ºC) to

process the materials (Meeker and Hamilton 2006).

It was necessary to grind rendering materials for transfer into stainless steel tubes.

Factors for comparing data to typical bone particle sizes will be necessary for future

experiments. Thermal conductivity studies on large bone particles could provide further

understanding of thermal lethality in rendering materials.

89

Figure 4.1. Method utilized to report RV and TT pre-enrichments results on XLD

confirmed by two confirmation tests at each thermal treatment. The result was assigned a

0 if both plates were negative (A). The result was assigned a 0.5 if one was positive and

one was negative (B). The result was assigned a 1.0 if both were positive (C).

A. B. C.

- - + - + +

0.0 0.5 1.0

90

Figure 4.2. Enumeration of Salmonella on XLD from poultry rendering samples (50%

fat) inoculated with Salmonella Choleraesuis (SC), Salmonella Enteriditis (SE),


1The lower limit of detection is 1.4 log10 cfu/g of Salmonella (n=24).

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Unheated

0 15

30

60

90

120

180

240

300

360

420

Mea

n B

acte

rial

Count,

log

10 c

fu/g


SC

SE

SN

SD

Inoculated

91

Figure 4.3. Enumeration of Salmonella on XLD from uninoculated poultry rendering




0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Unheated

0 15

30

60

90

120

180

240

300

360

420

Mea

n B

acte

rial

Count,

log

10 c

fu/g


SC

SE

SN

SD

Uninoculated

92

Figure 4.4. Enumeration of total bacteria on TSA from poultry rendering samples (50%

fat) inoculated with Salmonella Choleraesuis (SC), Salmonella Enteriditis (SE),



0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Unheated

0 15

30

60

90

120

180

240

300

360

420

Mea

n B

acte

rial

Count,

log

10 c

fu/g


SC

SE

SN

SD

Inoculated

93

Figure 4.5. Enumeration of total bacteria on TSA from uninoculated poultry rendering




0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Unheated

0 15

30

60

90

120

180

240

300

360

420

Mea

n B

acte

rial

Count,

log

10 c

fu/g


SC

SE

SN

SD

Uninoculated

94



Salmonella Dublin (SD) inoculated, RV pre-enriched poultry rendering samples (50%

fat).1

1A count of 0 represents the absence of Salmonella, while a count of 1 represents the

presence of Salmonella (n=24).

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Unheated

0 15

30

60

90

120

180

240

300

360

420

Pre

sence

/ A

bse

nce


SC

SE

SN

SD

Inoculated

95

Figure 4.7. Presence or absence ± standard deviation of Salmonella for each RV pre-

enriched, uninoculated poultry rendering samples (50% fat) for Salmonella Choleraesuis


(SD).1



0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Unheated

0 15

30

60

90

120

180

240

300

360

420

Pre

sence

/ A

bse

nce


SC

SE

SN

SD

Uninoculated

96



Salmonella Dublin (SD) inoculated, TT pre-enriched poultry rendering samples (50%

fat).1



0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Unheated

0 15

30

60

90

120

180

240

300

360

420

Pre

sence

/ A

bse

nce


SC

SE

SN

SD

Inoculated

97

Figure 4.9. Presence or absence ± standard deviation of Salmonella for each TT pre-

enriched, uninoculated poultry rendering samples (50% fat) for Salmonella Choleraesuis


(SD).1



0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Unheated

0 15

30

60

90

120

180

240

300

360

420

Pre

sence

/ A

bse

nce


SC

SE

SN

SD

Uninoculated

98

Table 4.1. Salmonella Choleraesuis (SC), Salmonella Enteriditis (SE), Salmonella

Newport (SN) and Salmonella Dublin (SD) after inoculation into poultry rendering

materials and plated onto XLD (n=6).

Serotype Average Broth Culture,


Average in Poultry Samples,


SC 12.60±0.15 10.47±0.20

SE 12.12±0.01 10.59±0.23

SN 12.28±0.03 10.43±0.22

SD 12.16±0.15 10.40±0.13

99


(SE), Salmonella Newport (SN), and Salmonella Dublin (SD) in poultry rendering

samples (50% fat) at 115.6ºC pre-enriched in RV and validated by two confirmation tests.


SC 0.60

SE 0.67

SN 0.39

SD 0.58

100


(SE), Salmonella Newport (SN), and Salmonella Dublin (SD) in poultry rendering

samples (50% fat) at 115.6ºC pre-enriched in TT and validated by two confirmation tests.


SC 0.70

SE 0.67

SN 0.67

SD 0.67

101

Table 4.4. Number of samples positive for Salmonella in Salmonella Choleraesuis (SC),

Salmonella Enteriditis (SE), Salmonella Newport (SN) and Salmonella Dublin (SD)

inoculated poultry rendering samples (50% fat) after pre-enrichment in RV or TT broth

(n=24).

RV TT

Serotype Thermal

Treatment

Time, s

Number of Positive

Samples

Number of Positive

Samples

































102

















103

Table 4.5. Number of samples positive for Salmonella in uninoculated poultry rendering


Salmonella Newport (SN) and Salmonella Dublin (SD) after pre-enrichment in RV or TT

broth (n=24).

RV TT

Serotype Thermal

Treatment

Time, s

Number of Positive

Samples

Number of Positive

Samples

















60 5 out of 24 samples 8 out of 24 samples 9















104


















105

References


R.M. 2011. Chapter 5 Salmonella. In FDA bacteriological analytical manual

(BAM). AOAC International, Gaithersburg, MD.


htm Accessed Apr. 2013.

Bucher, O., J.Y. D’Aoust, and R.A. Holley. 2008. Thermal resistance of Salmonella

serovars isolated from raw, frozen chicken nuggets/strips, nugget meat, and

pelleted broiled feed. Int. J. Food Microbiol. 124: 195–8.


and its relationship to human foodborne illness. Clin. Infect. Dis. 35: 859-865.

D’Aoust, J.Y. 2001. Foodborne salmonellosis: current international concerns. Food

Safety Mag. 7: 10-17.

Earle, R.L., and M.D. Earle. 1983. Unit operations in food processing. Web ed. The New

Zealand Institute of Food Science and Technology, Inc. New Zealand.

http://www.nzifst.org.nz/unitoperations/ Accessed Aug. 2013.

Feng, P. 2001. Rapid methods for detecting foodborne pathogens. In FDA bacteriological

analytical manual (BAM). AOAC International, Gaithersburg, MD.


htm Accessed Apr. 2013.

Franco, D.A. 1997. The genus Salmonella. Sanitation and hygiene in the production of

rendered animal by products. Animal Protein Producers Industry, National

Renderers Association, Fats and Proteins Research Foundation, Alexandria, VA.

Franco, D.A. 2005. A survey of Salmonella serovars and most probable numbers in

rendered-animal-protein meals: inferences for animal and human health. J.

Environ. Health. 67: 18-23.


poultry rendering materials. MS Thesis, Clemson Univ., Clemson. SC.

Hacking, W.C., W.R. Mitchell, and H.C. Carlson. 1977. Salmonella investigation in an

Ontario feed mill. Can. J. Comp. Med. 42: 400-406.

Heldman, D., and R. W. Hartel (ed.). 1998. Principles of food processing. Aspen

Publishers, Gaithersburg, MD.

106

Jay, J.M. 2005. Modern food microbiology, 6th ed. Aspen Publishers Inc., Gaithersburg,

MD.

Kinley, B. 2009. Prevalence and biological control of Salmonella contamination in

rendering plant environments and the finished rendered meals. PhD Diss.

Clemson Univ., Clemson, SC.

Kinley, B., J. Rieck, P. Dawson, and X. Jiang. 2010. Analysis of Salmonella and

enterococci isolated from rendered animal products. Can. J. Microbiol. 56: 65-73.

Laroussi, M., and F. Leipold. 2004. Evaluation of the roles of reactive species, heat, and

UV radiation in the inactivation of bacterial cells by air plasmas at atmospheric

pressure. Int. J. Mass Spectrom. 223: 81-86.

Loken, K.I., K.H. Culbert, R.E. Soles, and B.S. Pomeroy. 1968. Microbiological quality

of protein feed supplements produced by rendering plants. Appl. Microbiol. 16:

1002-1005.

Liu, T.S., G.H. Snoeyenbos, and V.L, Carlson. 1969. Thermal resistance of Salmonella

senftenberg 775W in dry animal feeds. Avian Dis. 13: 611-631.

McChesney, D.G., G. Kaplan, and P. Gardner. 1995. FDA survey determines Salmonella

contamination. Feedstuffs. 67: 20-23.

Meeker, D.L., and C.R. Hamilton. 2006. An overview of the rendering industry. Pages 1-

16 in Essential rendering all about the animal by-products industry. D.L. Meeker,

ed. National Renderers Association, Arlington, VA.

Murphy, R.Y., B.P. Marks, E.R. Johnson, and M.G. Johnson. 2000. Thermal inactivation

kinetics of Salmonella and Listeria in ground chicken breast meat and liquid

medium. J. Food Sci. 65: 706-710.

Murphy, R.Y., T. Osaili, L.K. Duncan, and J.A. Marcy. 2004. Thermal inactivation of

Salmonella and Listeria monocytogenes in ground chicken thigh/leg meat and

skin. Poult. Sci. 83: 1218-1225.

Ramirez-Lopez, L.M. 2006. Heat inactivation of thermo-resistant bacteria isolated from

poultry offal. MS Thesis. Clemson Univ., Clemson, SC.

Teixeira. A.A. 2006. Simulating thermal food processes using deterministic models.

Pages 73-106 in Thermal food processing. D. Sun, ed. CRC Press, Taylor &

Francis Group, Boca Raton, FL.

107

Troutt, H.F., D. Schaeffer, I. Kakoma, and G.G. Pearl. 2001. Prevalence of selected

foodborne pathogens in final rendered products. FPRF Directors Digest #312,

Fats and Proteins Research Foundation, Bloomington, IL.

Tuominen, L., T. Kairesalo, and H. Hartikainen. 1994. Comparison of methods for

Inhibiting bacterial activity in sediment. Appl Environ Microbiol.60: 3454–3457.


690.800 Salmonella in animal feed. Draft guidance. United States Food and Drug

Administration. Rockville, MD. http://www.gpo.gov/fdsys/pkg/FR-2010-08-

02/pdf/2010-18873.pdf Accessed May 2013.




policy-guide-sec-690800-salmonella Accessed Aug. 2013.

Date post:	27-Oct-2021
Category:	Documents
Upload:	others
View:	5 times
Download:	0 times